{
  "id": 551997,
  "title": "traindataset problem",
  "url": "/competitions/glioma-mcd-2025/discussion/551997",
  "author_name": "",
  "post_date": "2024-12-17T03:35:25.635415100Z",
  "votes": 3,
  "comment_count": 20,
  "views": 0,
  "content": "<p>The image size of some data in the training data is inconsistent with the image size in the json annotation.</p>",
  "messages": [
    {
      "id": "3073874",
      "postDate": "12/17/2024 03:35:25",
      "content": "<p>The image size of some data in the training data is inconsistent with the image size in the json annotation.</p>",
      "rawMarkdown": "The image size of some data in the training data is inconsistent with the image size in the json annotation.",
      "votes": null
    },
    {
      "id": "3074662",
      "postDate": "12/17/2024 22:59:21",
      "content": "<p>Thanks for pointing out Junqiang, we are tapping into the problem. Meanwhile, can you share the problematic data sample names you found?</p>",
      "rawMarkdown": "Thanks for pointing out Junqiang, we are tapping into the problem. Meanwhile, can you share the problematic data sample names you found?",
      "votes": null
    },
    {
      "id": "3074747",
      "postDate": "12/18/2024 02:50:39",
      "content": "<p>the train dataset,i find imagesize is 512x512,but the related json file is 1015x1015<br>\n(1015, 1015) (512, 512)<br>\ntraining0002.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0003.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0004.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0005.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0006.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0007.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0008.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0009.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0010.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0011.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0012.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0013.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0014.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0015.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0016.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0017.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0018.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0019.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0020.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0021.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0022.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0023.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0024.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0025.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0026.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0027.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0028.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0029.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0030.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0031.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0032.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0033.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0034.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0035.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0036.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0037.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0038.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0039.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0040.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0041.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0042.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0043.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0044.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0045.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0046.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0047.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0048.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0049.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0050.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0051.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0052.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0053.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0054.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0055.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0056.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0057.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0058.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0059.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0060.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0061.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0062.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0063.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0064.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0065.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0066.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0067.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0068.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0069.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0070.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0071.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0072.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0073.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0074.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0075.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0076.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0077.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0078.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0079.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0080.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0081.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0082.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0083.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0084.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0085.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0086.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0087.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0088.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0089.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0090.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0091.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0092.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0093.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0094.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0095.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0096.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0097.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0098.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0099.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0100.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0101.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0102.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0103.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0104.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0105.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0106.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0107.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0108.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0109.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0110.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0111.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0112.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0113.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0114.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0115.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0116.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0117.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0118.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0119.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0120.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0121.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0122.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0123.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0124.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0125.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0126.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0127.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0128.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0129.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0130.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0131.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0132.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0133.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0134.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0135.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0136.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0137.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0138.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0139.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0140.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0141.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0142.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0143.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0144.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0145.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0146.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0147.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0148.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0149.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0150.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0151.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0152.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0153.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0154.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0155.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0156.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0157.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0158.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0159.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0160.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0161.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0162.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0163.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0164.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0165.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0166.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0167.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0168.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0169.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0170.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0171.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0172.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0173.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0174.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0175.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0176.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0177.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0178.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0179.json</p>",
      "rawMarkdown": "the train dataset,i find imagesize is 512x512,but the related json file is 1015x1015\n(1015, 1015) (512, 512)\ntraining0002.json\n(1015, 1015) (512, 512)\ntraining0003.json\n(1015, 1015) (512, 512)\ntraining0004.json\n(1015, 1015) (512, 512)\ntraining0005.json\n(1015, 1015) (512, 512)\ntraining0006.json\n(1015, 1015) (512, 512)\ntraining0007.json\n(1015, 1015) (512, 512)\ntraining0008.json\n(1015, 1015) (512, 512)\ntraining0009.json\n(1015, 1015) (512, 512)\ntraining0010.json\n(1015, 1015) (512, 512)\ntraining0011.json\n(1015, 1015) (512, 512)\ntraining0012.json\n(1015, 1015) (512, 512)\ntraining0013.json\n(1015, 1015) (512, 512)\ntraining0014.json\n(1015, 1015) (512, 512)\ntraining0015.json\n(1015, 1015) (512, 512)\ntraining0016.json\n(1015, 1015) (512, 512)\ntraining0017.json\n(1015, 1015) (512, 512)\ntraining0018.json\n(1015, 1015) (512, 512)\ntraining0019.json\n(1015, 1015) (512, 512)\ntraining0020.json\n(1015, 1015) (512, 512)\ntraining0021.json\n(1015, 1015) (512, 512)\ntraining0022.json\n(1015, 1015) (512, 512)\ntraining0023.json\n(1015, 1015) (512, 512)\ntraining0024.json\n(1015, 1015) (512, 512)\ntraining0025.json\n(1015, 1015) (512, 512)\ntraining0026.json\n(1015, 1015) (512, 512)\ntraining0027.json\n(1015, 1015) (512, 512)\ntraining0028.json\n(1015, 1015) (512, 512)\ntraining0029.json\n(1015, 1015) (512, 512)\ntraining0030.json\n(1015, 1015) (512, 512)\ntraining0031.json\n(1015, 1015) (512, 512)\ntraining0032.json\n(1015, 1015) (512, 512)\ntraining0033.json\n(1015, 1015) (512, 512)\ntraining0034.json\n(1015, 1015) (512, 512)\ntraining0035.json\n(1015, 1015) (512, 512)\ntraining0036.json\n(1015, 1015) (512, 512)\ntraining0037.json\n(1015, 1015) (512, 512)\ntraining0038.json\n(1015, 1015) (512, 512)\ntraining0039.json\n(1015, 1015) (512, 512)\ntraining0040.json\n(1015, 1015) (512, 512)\ntraining0041.json\n(1015, 1015) (512, 512)\ntraining0042.json\n(1015, 1015) (512, 512)\ntraining0043.json\n(1015, 1015) (512, 512)\ntraining0044.json\n(1015, 1015) (512, 512)\ntraining0045.json\n(1015, 1015) (512, 512)\ntraining0046.json\n(1015, 1015) (512, 512)\ntraining0047.json\n(1015, 1015) (512, 512)\ntraining0048.json\n(1015, 1015) (512, 512)\ntraining0049.json\n(1015, 1015) (512, 512)\ntraining0050.json\n(1015, 1015) (512, 512)\ntraining0051.json\n(1015, 1015) (512, 512)\ntraining0052.json\n(1015, 1015) (512, 512)\ntraining0053.json\n(1015, 1015) (512, 512)\ntraining0054.json\n(1015, 1015) (512, 512)\ntraining0055.json\n(1015, 1015) (512, 512)\ntraining0056.json\n(1015, 1015) (512, 512)\ntraining0057.json\n(1015, 1015) (512, 512)\ntraining0058.json\n(1015, 1015) (512, 512)\ntraining0059.json\n(1015, 1015) (512, 512)\ntraining0060.json\n(1015, 1015) (512, 512)\ntraining0061.json\n(1015, 1015) (512, 512)\ntraining0062.json\n(1015, 1015) (512, 512)\ntraining0063.json\n(1015, 1015) (512, 512)\ntraining0064.json\n(1015, 1015) (512, 512)\ntraining0065.json\n(1015, 1015) (512, 512)\ntraining0066.json\n(1015, 1015) (512, 512)\ntraining0067.json\n(1015, 1015) (512, 512)\ntraining0068.json\n(1015, 1015) (512, 512)\ntraining0069.json\n(1015, 1015) (512, 512)\ntraining0070.json\n(1015, 1015) (512, 512)\ntraining0071.json\n(1015, 1015) (512, 512)\ntraining0072.json\n(1015, 1015) (512, 512)\ntraining0073.json\n(1015, 1015) (512, 512)\ntraining0074.json\n(1015, 1015) (512, 512)\ntraining0075.json\n(1015, 1015) (512, 512)\ntraining0076.json\n(1015, 1015) (512, 512)\ntraining0077.json\n(1015, 1015) (512, 512)\ntraining0078.json\n(1015, 1015) (512, 512)\ntraining0079.json\n(1015, 1015) (512, 512)\ntraining0080.json\n(1015, 1015) (512, 512)\ntraining0081.json\n(1015, 1015) (512, 512)\ntraining0082.json\n(1015, 1015) (512, 512)\ntraining0083.json\n(1015, 1015) (512, 512)\ntraining0084.json\n(1015, 1015) (512, 512)\ntraining0085.json\n(1015, 1015) (512, 512)\ntraining0086.json\n(1015, 1015) (512, 512)\ntraining0087.json\n(1015, 1015) (512, 512)\ntraining0088.json\n(1015, 1015) (512, 512)\ntraining0089.json\n(1015, 1015) (512, 512)\ntraining0090.json\n(1015, 1015) (512, 512)\ntraining0091.json\n(1015, 1015) (512, 512)\ntraining0092.json\n(1015, 1015) (512, 512)\ntraining0093.json\n(1015, 1015) (512, 512)\ntraining0094.json\n(1015, 1015) (512, 512)\ntraining0095.json\n(1015, 1015) (512, 512)\ntraining0096.json\n(1015, 1015) (512, 512)\ntraining0097.json\n(1015, 1015) (512, 512)\ntraining0098.json\n(1015, 1015) (512, 512)\ntraining0099.json\n(1015, 1015) (512, 512)\ntraining0100.json\n(1015, 1015) (512, 512)\ntraining0101.json\n(1015, 1015) (512, 512)\ntraining0102.json\n(1015, 1015) (512, 512)\ntraining0103.json\n(1015, 1015) (512, 512)\ntraining0104.json\n(1015, 1015) (512, 512)\ntraining0105.json\n(1015, 1015) (512, 512)\ntraining0106.json\n(1015, 1015) (512, 512)\ntraining0107.json\n(1015, 1015) (512, 512)\ntraining0108.json\n(1015, 1015) (512, 512)\ntraining0109.json\n(1015, 1015) (512, 512)\ntraining0110.json\n(1015, 1015) (512, 512)\ntraining0111.json\n(1015, 1015) (512, 512)\ntraining0112.json\n(1015, 1015) (512, 512)\ntraining0113.json\n(1015, 1015) (512, 512)\ntraining0114.json\n(1015, 1015) (512, 512)\ntraining0115.json\n(1015, 1015) (512, 512)\ntraining0116.json\n(1015, 1015) (512, 512)\ntraining0117.json\n(1015, 1015) (512, 512)\ntraining0118.json\n(1015, 1015) (512, 512)\ntraining0119.json\n(1015, 1015) (512, 512)\ntraining0120.json\n(1015, 1015) (512, 512)\ntraining0121.json\n(1015, 1015) (512, 512)\ntraining0122.json\n(1015, 1015) (512, 512)\ntraining0123.json\n(1015, 1015) (512, 512)\ntraining0124.json\n(1015, 1015) (512, 512)\ntraining0125.json\n(1015, 1015) (512, 512)\ntraining0126.json\n(1015, 1015) (512, 512)\ntraining0127.json\n(1015, 1015) (512, 512)\ntraining0128.json\n(1015, 1015) (512, 512)\ntraining0129.json\n(1015, 1015) (512, 512)\ntraining0130.json\n(1015, 1015) (512, 512)\ntraining0131.json\n(1015, 1015) (512, 512)\ntraining0132.json\n(1015, 1015) (512, 512)\ntraining0133.json\n(1015, 1015) (512, 512)\ntraining0134.json\n(1015, 1015) (512, 512)\ntraining0135.json\n(1015, 1015) (512, 512)\ntraining0136.json\n(1015, 1015) (512, 512)\ntraining0137.json\n(1015, 1015) (512, 512)\ntraining0138.json\n(1015, 1015) (512, 512)\ntraining0139.json\n(1015, 1015) (512, 512)\ntraining0140.json\n(1015, 1015) (512, 512)\ntraining0141.json\n(1015, 1015) (512, 512)\ntraining0142.json\n(1015, 1015) (512, 512)\ntraining0143.json\n(1015, 1015) (512, 512)\ntraining0144.json\n(1015, 1015) (512, 512)\ntraining0145.json\n(1015, 1015) (512, 512)\ntraining0146.json\n(1015, 1015) (512, 512)\ntraining0147.json\n(1015, 1015) (512, 512)\ntraining0148.json\n(1015, 1015) (512, 512)\ntraining0149.json\n(1015, 1015) (512, 512)\ntraining0150.json\n(1015, 1015) (512, 512)\ntraining0151.json\n(1015, 1015) (512, 512)\ntraining0152.json\n(1015, 1015) (512, 512)\ntraining0153.json\n(1015, 1015) (512, 512)\ntraining0154.json\n(1015, 1015) (512, 512)\ntraining0155.json\n(1015, 1015) (512, 512)\ntraining0156.json\n(1015, 1015) (512, 512)\ntraining0157.json\n(1015, 1015) (512, 512)\ntraining0158.json\n(1015, 1015) (512, 512)\ntraining0159.json\n(1015, 1015) (512, 512)\ntraining0160.json\n(1015, 1015) (512, 512)\ntraining0161.json\n(1015, 1015) (512, 512)\ntraining0162.json\n(1015, 1015) (512, 512)\ntraining0163.json\n(1015, 1015) (512, 512)\ntraining0164.json\n(1015, 1015) (512, 512)\ntraining0165.json\n(1015, 1015) (512, 512)\ntraining0166.json\n(1015, 1015) (512, 512)\ntraining0167.json\n(1015, 1015) (512, 512)\ntraining0168.json\n(1015, 1015) (512, 512)\ntraining0169.json\n(1015, 1015) (512, 512)\ntraining0170.json\n(1015, 1015) (512, 512)\ntraining0171.json\n(1015, 1015) (512, 512)\ntraining0172.json\n(1015, 1015) (512, 512)\ntraining0173.json\n(1015, 1015) (512, 512)\ntraining0174.json\n(1015, 1015) (512, 512)\ntraining0175.json\n(1015, 1015) (512, 512)\ntraining0176.json\n(1015, 1015) (512, 512)\ntraining0177.json\n(1015, 1015) (512, 512)\ntraining0178.json\n(1015, 1015) (512, 512)\ntraining0179.json",
      "votes": null
    },
    {
      "id": "3074749",
      "postDate": "12/18/2024 02:51:39",
      "content": "<p>the test dataset,i find imagesize is 512x512,but the related json file is 1015x1015<br>\n(1015, 1015) (512, 512)<br>\ntesting0002.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0003.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0004.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0005.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0006.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0007.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0008.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0009.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0010.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0011.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0012.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0013.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0014.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0015.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0016.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0017.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0018.json</p>",
      "rawMarkdown": "the test dataset,i find imagesize is 512x512,but the related json file is 1015x1015\n(1015, 1015) (512, 512)\ntesting0002.json\n(1015, 1015) (512, 512)\ntesting0003.json\n(1015, 1015) (512, 512)\ntesting0004.json\n(1015, 1015) (512, 512)\ntesting0005.json\n(1015, 1015) (512, 512)\ntesting0006.json\n(1015, 1015) (512, 512)\ntesting0007.json\n(1015, 1015) (512, 512)\ntesting0008.json\n(1015, 1015) (512, 512)\ntesting0009.json\n(1015, 1015) (512, 512)\ntesting0010.json\n(1015, 1015) (512, 512)\ntesting0011.json\n(1015, 1015) (512, 512)\ntesting0012.json\n(1015, 1015) (512, 512)\ntesting0013.json\n(1015, 1015) (512, 512)\ntesting0014.json\n(1015, 1015) (512, 512)\ntesting0015.json\n(1015, 1015) (512, 512)\ntesting0016.json\n(1015, 1015) (512, 512)\ntesting0017.json\n(1015, 1015) (512, 512)\ntesting0018.json",
      "votes": null
    },
    {
      "id": "3075502",
      "postDate": "12/19/2024 00:35:57",
      "content": "<p>Thanks for the time to find out the unmatchable files. We have come up with a solution to solve the unmatching problem. Please use this code below to <em>up-sample</em> the 512-by-512 images to 1015-by-1015 dimensions. Let me know if you need any further help.</p>\n<pre><code> cv2\n os\n\n\ninput_dir = \noutput_dir = \n\n\n  os.path.exists(output_dir):\n    os.makedirs(output_dir)\n\n\ntarget_size = (, )\n\n\nimage_files = [f  f  os.listdir(input_dir)  f.lower().endswith()]\n\n\n img_name  image_files:\n    input_path = os.path.join(input_dir, img_name)\n    output_path = os.path.join(output_dir, img_name.replace(, ))\n\n    \n    img = cv2.imread(input_path)\n\n     img  :\n        ()\n        \n\n    \n    resized_img = cv2.resize(img, target_size, interpolation=cv2.INTER_CUBIC)\n\n    \n    cv2.imwrite(output_path, resized_img)\n    ()\n</code></pre>",
      "rawMarkdown": "Thanks for the time to find out the unmatchable files. We have come up with a solution to solve the unmatching problem. Please use this code below to *up-sample* the 512-by-512 images to 1015-by-1015 dimensions. Let me know if you need any further help.\n\n\n\n    import cv2\n    import os\n\n    # Set the input and output folder paths\n    input_dir = r\"input_images\"\n    output_dir = r\"output_images\"\n\n    # If the output folder does not exist, create it\n    if not os.path.exists(output_dir):\n        os.makedirs(output_dir)\n\n    # Target size (1015x1015)\n    target_size = (1015, 1015)\n\n    # Get all .jpg image files from the input folder\n    image_files = [f for f in os.listdir(input_dir) if f.lower().endswith('.jpg')]\n\n    # Batch upsampling\n    for img_name in image_files:\n        input_path = os.path.join(input_dir, img_name)\n        output_path = os.path.join(output_dir, img_name.replace('.jpg', '_resized.jpg'))\n\n        # Read the original image\n        img = cv2.imread(input_path)\n        \n        if img is None:\n            print(f\"Unable to read image: {input_path}, skipping...\")\n            continue\n        \n        # Use bicubic interpolation for upsampling\n        resized_img = cv2.resize(img, target_size, interpolation=cv2.INTER_CUBIC)\n        \n        # Save the processed image\n        cv2.imwrite(output_path, resized_img)\n        print(f\"Processing completed: {output_path}\")",
      "votes": null
    },
    {
      "id": "3075559",
      "postDate": "12/19/2024 02:43:52",
      "content": "<p>thank you,it's working</p>",
      "rawMarkdown": "thank you,it's working",
      "votes": null
    },
    {
      "id": "3081223",
      "postDate": "12/26/2024 13:02:31",
      "content": "<p>I think there is some issue with the JSON  files. I was checking the annotations of file number \"training0016.json\" via labelme, but the image it shows is actually \"training0015.jpg\". To verify, I double-checked it by plotting it using Python. I am assuming the numbering of JSON is wrong.</p>",
      "rawMarkdown": "I think there is some issue with the JSON  files. I was checking the annotations of file number \"training0016.json\" via labelme, but the image it shows is actually \"training0015.jpg\". To verify, I double-checked it by plotting it using Python. I am assuming the numbering of JSON is wrong.",
      "votes": null
    },
    {
      "id": "3082027",
      "postDate": "12/27/2024 14:24:08",
      "content": "<p>Sorry for the repeated problems with this dataset, and thank you for pointing out the unmatch issue. After discussion, we will redo the annotation on the 1015-by-1015 samples; please give us two to three days to finish. While waiting for the updated version, you can still train your model based on 512-by-512 samples, those are not going to be re-annotated. Thank you again, and don't hesitate to get in touch with us if you have any further questions. Happy holiday ;-)</p>",
      "rawMarkdown": "Sorry for the repeated problems with this dataset, and thank you for pointing out the unmatch issue. After discussion, we will redo the annotation on the 1015-by-1015 samples; please give us two to three days to finish. While waiting for the updated version, you can still train your model based on 512-by-512 samples, those are not going to be re-annotated. Thank you again, and don't hesitate to get in touch with us if you have any further questions. Happy holiday ;-)",
      "votes": null
    },
    {
      "id": "3082937",
      "postDate": "12/28/2024 19:30:50",
      "content": "<p>Thanks all for your engagement.</p>\n<p>We have updated our dataset to version <strong>Data_122824</strong>. Please use the most up-to-date version to train your model.<br>\nWhat has been changed in this version:</p>\n<ol>\n<li>Re-annotated unclear images for better training results.</li>\n<li>Fixed the unmatched JSON and JPG files</li>\n<li>Added one more test sample (you may need to use the most up-to-date Submission_template.csv)</li>\n</ol>\n<p><strong>The data format and evaluation metric remained the same. You don't have to worry about changing your model structure.</strong></p>",
      "rawMarkdown": "Thanks all for your engagement.\n\nWe have updated our dataset to version **Data_122824**. Please use the most up-to-date version to train your model.\nWhat has been changed in this version:\n\n1. Re-annotated unclear images for better training results.\n2. Fixed the unmatched JSON and JPG files\n3. Added one more test sample (you may need to use the most up-to-date Submission_template.csv)\n\n**The data format and evaluation metric remained the same. You don't have to worry about changing your model structure.**",
      "votes": null
    },
    {
      "id": "3088132",
      "postDate": "01/04/2025 11:18:12",
      "content": "<p>I'm using Data_122824 but yet even training0001.jpg seems to have a problem. The image size is 512x512 but JSON coordinates include larger than 512, indicating it's actually 1015. </p>",
      "rawMarkdown": "I'm using Data_122824 but yet even training0001.jpg seems to have a problem. The image size is 512x512 but JSON coordinates include larger than 512, indicating it's actually 1015.",
      "votes": null
    },
    {
      "id": "3088687",
      "postDate": "01/05/2025 03:03:04",
      "content": "<p>Thanks, we are looking into it.</p>",
      "rawMarkdown": "Thanks, we are looking into it.",
      "votes": null
    },
    {
      "id": "3090770",
      "postDate": "01/07/2025 16:46:30",
      "content": "<p>Thanks for waiting, after discussion with our team, please use the code in this question to preprocess the dimension.</p>",
      "rawMarkdown": "Thanks for waiting, after discussion with our team, please use the code in this question to preprocess the dimension.",
      "votes": null
    },
    {
      "id": "3092394",
      "postDate": "01/09/2025 13:59:26",
      "content": "<p>excuseme, I find the wbsite data explorer are correct shape ,but not equals the website when i download , for example training0003.json in Data_122824 folder</p>",
      "rawMarkdown": "excuseme, I find the wbsite data explorer are correct shape ,but not equals the website when i download , for example training0003.json in Data_122824 folder",
      "votes": null
    },
    {
      "id": "3096317",
      "postDate": "01/14/2025 08:44:46",
      "content": "<p>Hi,</p>\n<p>I also encountered an issue with unmatched JSON and JPG files in the latest updated dataset, Data_122824.</p>\n<p>For example, the image \"training0010.jpg\" in the left side of Fig. 1 is the original version, but after loading it into LabelMe, it displays a different case. This discrepancy may lead to annotation errors when visualizing the labels from the corresponding JSON files in the original image (as shown in Fig. 2). </p>\n<p>I’ve found that this problem affects several other files, including \"training0010.jpg,\" \"training0080.jpg,\" \"training0099.jpg,\" \"training0102.jpg,\" \"training0114.jpg,\" and others.</p>\n<p>Has anyone else encountered and solved this problem? Or could you advise on how I can fix the JSON files?</p>\n<p>Thank you for your assistance.</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4596430%2Fc6288be968947b9d066e23860fa68611%2F1.png?generation=1736844271316451&amp;alt=media\" alt=\"\"><br>\nfig1. the original image and showed in lableme</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4596430%2F7cdb7f2cf4de06b95f77eaedeb4c780f%2F2.png?generation=1736844283864866&amp;alt=media\" alt=\"\">  <br>\nfig2. the visualized image (case: training0010.jpg)</p>",
      "rawMarkdown": "Hi,\n\nI also encountered an issue with unmatched JSON and JPG files in the latest updated dataset, Data_122824.\n\nFor example, the image \"training0010.jpg\" in the left side of Fig. 1 is the original version, but after loading it into LabelMe, it displays a different case. This discrepancy may lead to annotation errors when visualizing the labels from the corresponding JSON files in the original image (as shown in Fig. 2). \n\nI’ve found that this problem affects several other files, including \"training0010.jpg,\" \"training0080.jpg,\" \"training0099.jpg,\" \"training0102.jpg,\" \"training0114.jpg,\" and others.\n\nHas anyone else encountered and solved this problem? Or could you advise on how I can fix the JSON files?\n\nThank you for your assistance.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4596430%2Fc6288be968947b9d066e23860fa68611%2F1.png?generation=1736844271316451&alt=media)\nfig1. the original image and showed in lableme\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4596430%2F7cdb7f2cf4de06b95f77eaedeb4c780f%2F2.png?generation=1736844283864866&alt=media)  \nfig2. the visualized image (case: training0010.jpg)",
      "votes": null
    },
    {
      "id": "3097256",
      "postDate": "01/15/2025 07:37:05",
      "content": "<p>Dear <a href=\"https://www.kaggle.com/wentaolimda\" target=\"_blank\">@wentaolimda</a> Is there any solution i missed or you are still progress this issue? Thank you very much</p>",
      "rawMarkdown": "Dear @wentaolimda Is there any solution i missed or you are still progress this issue? Thank you very much",
      "votes": null
    },
    {
      "id": "3097713",
      "postDate": "01/15/2025 16:26:40",
      "content": "<p>Sorry for the late reply as our team has been working continuously on this problem. Right now, we don't know why exactly the lableme shows different images (actually it always pointed to training0001 from the problematic ones). While we are still working on the problem, please ignore these images for now. </p>",
      "rawMarkdown": "Sorry for the late reply as our team has been working continuously on this problem. Right now, we don't know why exactly the lableme shows different images (actually it always pointed to training0001 from the problematic ones). While we are still working on the problem, please ignore these images for now.",
      "votes": null
    },
    {
      "id": "3098078",
      "postDate": "01/16/2025 03:44:54",
      "content": "<p>Thank you for your reply. I understand now.</p>",
      "rawMarkdown": "Thank you for your reply. I understand now.",
      "votes": null
    },
    {
      "id": "3101061",
      "postDate": "01/20/2025 09:54:31",
      "content": "<p>Today I joined the competition,  so probably I use the most updated train dataset. Still issue with coordination. train image 0002 has two polygons both Non-mitosis one of them in rage 512*512 but the second polygon far beyond </p>",
      "rawMarkdown": "Today I joined the competition,  so probably I use the most updated train dataset. Still issue with coordination. train image 0002 has two polygons both Non-mitosis one of them in rage 512*512 but the second polygon far beyond",
      "votes": null
    },
    {
      "id": "3101063",
      "postDate": "01/20/2025 09:57:17",
      "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F16163090%2Fb2b5feacebdcee5cf0c73905fa7761ac%2FScreenshot%202025-01-20%20115700.png?generation=1737367032272456&amp;alt=media\" alt=\"\"></p>",
      "rawMarkdown": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F16163090%2Fb2b5feacebdcee5cf0c73905fa7761ac%2FScreenshot%202025-01-20%20115700.png?generation=1737367032272456&alt=media)",
      "votes": null
    },
    {
      "id": "3105905",
      "postDate": "01/24/2025 05:43:41",
      "content": "<p>I've tried to decode the image_data in the JSON of training0002. It’s a 1015x1015 image, so I think this JSON provides the wrong value for imageHeight</p>",
      "rawMarkdown": "I've tried to decode the image_data in the JSON of training0002. It’s a 1015x1015 image, so I think this JSON provides the wrong value for imageHeight",
      "votes": null
    },
    {
      "id": "3194818",
      "postDate": "05/06/2025 10:35:21",
      "content": "<p>Hi!<br>\nI have trained some models and I'm having problems regarding the score of the test set. My models get satisfactory results on train and validation sets, but when predicting the test set, the performance is the same as a random one. I also created a test set from the train data to check for overfitting, and the results were similar to validation, so I don’t think it’s overfitting. Has anyone had this issue? Does anyone know why this might be happening?</p>",
      "rawMarkdown": "Hi!\nI have trained some models and I'm having problems regarding the score of the test set. My models get satisfactory results on train and validation sets, but when predicting the test set, the performance is the same as a random one. I also created a test set from the train data to check for overfitting, and the results were similar to validation, so I don’t think it’s overfitting. Has anyone had this issue? Does anyone know why this might be happening?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3074662,
      "author_name": "wentaolimda",
      "author_url": "",
      "post_date": "12/17/2024 22:59:21",
      "content": "<p>Thanks for pointing out Junqiang, we are tapping into the problem. Meanwhile, can you share the problematic data sample names you found?</p>",
      "votes": null,
      "replies": [
        {
          "id": 3074747,
          "author_name": "junqiangchen90",
          "author_url": "",
          "post_date": "12/18/2024 02:50:39",
          "content": "<p>the train dataset,i find imagesize is 512x512,but the related json file is 1015x1015<br>\n(1015, 1015) (512, 512)<br>\ntraining0002.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0003.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0004.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0005.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0006.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0007.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0008.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0009.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0010.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0011.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0012.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0013.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0014.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0015.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0016.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0017.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0018.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0019.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0020.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0021.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0022.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0023.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0024.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0025.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0026.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0027.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0028.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0029.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0030.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0031.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0032.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0033.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0034.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0035.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0036.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0037.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0038.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0039.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0040.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0041.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0042.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0043.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0044.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0045.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0046.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0047.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0048.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0049.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0050.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0051.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0052.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0053.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0054.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0055.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0056.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0057.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0058.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0059.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0060.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0061.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0062.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0063.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0064.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0065.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0066.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0067.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0068.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0069.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0070.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0071.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0072.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0073.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0074.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0075.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0076.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0077.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0078.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0079.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0080.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0081.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0082.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0083.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0084.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0085.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0086.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0087.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0088.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0089.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0090.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0091.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0092.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0093.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0094.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0095.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0096.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0097.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0098.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0099.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0100.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0101.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0102.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0103.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0104.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0105.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0106.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0107.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0108.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0109.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0110.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0111.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0112.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0113.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0114.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0115.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0116.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0117.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0118.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0119.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0120.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0121.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0122.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0123.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0124.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0125.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0126.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0127.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0128.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0129.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0130.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0131.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0132.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0133.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0134.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0135.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0136.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0137.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0138.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0139.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0140.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0141.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0142.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0143.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0144.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0145.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0146.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0147.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0148.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0149.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0150.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0151.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0152.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0153.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0154.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0155.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0156.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0157.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0158.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0159.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0160.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0161.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0162.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0163.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0164.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0165.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0166.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0167.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0168.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0169.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0170.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0171.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0172.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0173.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0174.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0175.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0176.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0177.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0178.json<br>\n(1015, 1015) (512, 512)<br>\ntraining0179.json</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 3074749,
          "author_name": "junqiangchen90",
          "author_url": "",
          "post_date": "12/18/2024 02:51:39",
          "content": "<p>the test dataset,i find imagesize is 512x512,but the related json file is 1015x1015<br>\n(1015, 1015) (512, 512)<br>\ntesting0002.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0003.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0004.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0005.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0006.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0007.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0008.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0009.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0010.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0011.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0012.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0013.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0014.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0015.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0016.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0017.json<br>\n(1015, 1015) (512, 512)<br>\ntesting0018.json</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 3075502,
      "author_name": "wentaolimda",
      "author_url": "",
      "post_date": "12/19/2024 00:35:57",
      "content": "<p>Thanks for the time to find out the unmatchable files. We have come up with a solution to solve the unmatching problem. Please use this code below to <em>up-sample</em> the 512-by-512 images to 1015-by-1015 dimensions. Let me know if you need any further help.</p>\n<pre><code> cv2\n os\n\n\ninput_dir = \noutput_dir = \n\n\n  os.path.exists(output_dir):\n    os.makedirs(output_dir)\n\n\ntarget_size = (, )\n\n\nimage_files = [f  f  os.listdir(input_dir)  f.lower().endswith()]\n\n\n img_name  image_files:\n    input_path = os.path.join(input_dir, img_name)\n    output_path = os.path.join(output_dir, img_name.replace(, ))\n\n    \n    img = cv2.imread(input_path)\n\n     img  :\n        ()\n        \n\n    \n    resized_img = cv2.resize(img, target_size, interpolation=cv2.INTER_CUBIC)\n\n    \n    cv2.imwrite(output_path, resized_img)\n    ()\n</code></pre>",
      "votes": null,
      "replies": [
        {
          "id": 3075559,
          "author_name": "junqiangchen90",
          "author_url": "",
          "post_date": "12/19/2024 02:43:52",
          "content": "<p>thank you,it's working</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 3081223,
      "author_name": "usmabhatt",
      "author_url": "",
      "post_date": "12/26/2024 13:02:31",
      "content": "<p>I think there is some issue with the JSON  files. I was checking the annotations of file number \"training0016.json\" via labelme, but the image it shows is actually \"training0015.jpg\". To verify, I double-checked it by plotting it using Python. I am assuming the numbering of JSON is wrong.</p>",
      "votes": null,
      "replies": [
        {
          "id": 3082027,
          "author_name": "wentaolimda",
          "author_url": "",
          "post_date": "12/27/2024 14:24:08",
          "content": "<p>Sorry for the repeated problems with this dataset, and thank you for pointing out the unmatch issue. After discussion, we will redo the annotation on the 1015-by-1015 samples; please give us two to three days to finish. While waiting for the updated version, you can still train your model based on 512-by-512 samples, those are not going to be re-annotated. Thank you again, and don't hesitate to get in touch with us if you have any further questions. Happy holiday ;-)</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 3082937,
      "author_name": "wentaolimda",
      "author_url": "",
      "post_date": "12/28/2024 19:30:50",
      "content": "<p>Thanks all for your engagement.</p>\n<p>We have updated our dataset to version <strong>Data_122824</strong>. Please use the most up-to-date version to train your model.<br>\nWhat has been changed in this version:</p>\n<ol>\n<li>Re-annotated unclear images for better training results.</li>\n<li>Fixed the unmatched JSON and JPG files</li>\n<li>Added one more test sample (you may need to use the most up-to-date Submission_template.csv)</li>\n</ol>\n<p><strong>The data format and evaluation metric remained the same. You don't have to worry about changing your model structure.</strong></p>",
      "votes": null,
      "replies": [
        {
          "id": 3088132,
          "author_name": "justinkim1216",
          "author_url": "",
          "post_date": "01/04/2025 11:18:12",
          "content": "<p>I'm using Data_122824 but yet even training0001.jpg seems to have a problem. The image size is 512x512 but JSON coordinates include larger than 512, indicating it's actually 1015. </p>",
          "votes": null,
          "replies": [
            {
              "id": 3088687,
              "author_name": "wentaolimda",
              "author_url": "",
              "post_date": "01/05/2025 03:03:04",
              "content": "<p>Thanks, we are looking into it.</p>",
              "votes": null,
              "replies": []
            },
            {
              "id": 3090770,
              "author_name": "wentaolimda",
              "author_url": "",
              "post_date": "01/07/2025 16:46:30",
              "content": "<p>Thanks for waiting, after discussion with our team, please use the code in this question to preprocess the dimension.</p>",
              "votes": null,
              "replies": [
                {
                  "id": 3092394,
                  "author_name": "hsmy955",
                  "author_url": "",
                  "post_date": "01/09/2025 13:59:26",
                  "content": "<p>excuseme, I find the wbsite data explorer are correct shape ,but not equals the website when i download , for example training0003.json in Data_122824 folder</p>",
                  "votes": null,
                  "replies": []
                }
              ]
            }
          ]
        }
      ]
    },
    {
      "id": 3096317,
      "author_name": "chingpingwang",
      "author_url": "",
      "post_date": "01/14/2025 08:44:46",
      "content": "<p>Hi,</p>\n<p>I also encountered an issue with unmatched JSON and JPG files in the latest updated dataset, Data_122824.</p>\n<p>For example, the image \"training0010.jpg\" in the left side of Fig. 1 is the original version, but after loading it into LabelMe, it displays a different case. This discrepancy may lead to annotation errors when visualizing the labels from the corresponding JSON files in the original image (as shown in Fig. 2). </p>\n<p>I’ve found that this problem affects several other files, including \"training0010.jpg,\" \"training0080.jpg,\" \"training0099.jpg,\" \"training0102.jpg,\" \"training0114.jpg,\" and others.</p>\n<p>Has anyone else encountered and solved this problem? Or could you advise on how I can fix the JSON files?</p>\n<p>Thank you for your assistance.</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4596430%2Fc6288be968947b9d066e23860fa68611%2F1.png?generation=1736844271316451&amp;alt=media\" alt=\"\"><br>\nfig1. the original image and showed in lableme</p>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4596430%2F7cdb7f2cf4de06b95f77eaedeb4c780f%2F2.png?generation=1736844283864866&amp;alt=media\" alt=\"\">  <br>\nfig2. the visualized image (case: training0010.jpg)</p>",
      "votes": null,
      "replies": [
        {
          "id": 3097256,
          "author_name": "chingpingwang",
          "author_url": "",
          "post_date": "01/15/2025 07:37:05",
          "content": "<p>Dear <a href=\"https://www.kaggle.com/wentaolimda\" target=\"_blank\">@wentaolimda</a> Is there any solution i missed or you are still progress this issue? Thank you very much</p>",
          "votes": null,
          "replies": [
            {
              "id": 3097713,
              "author_name": "wentaolimda",
              "author_url": "",
              "post_date": "01/15/2025 16:26:40",
              "content": "<p>Sorry for the late reply as our team has been working continuously on this problem. Right now, we don't know why exactly the lableme shows different images (actually it always pointed to training0001 from the problematic ones). While we are still working on the problem, please ignore these images for now. </p>",
              "votes": null,
              "replies": [
                {
                  "id": 3098078,
                  "author_name": "chingpingwang",
                  "author_url": "",
                  "post_date": "01/16/2025 03:44:54",
                  "content": "<p>Thank you for your reply. I understand now.</p>",
                  "votes": null,
                  "replies": []
                }
              ]
            }
          ]
        }
      ]
    },
    {
      "id": 3101061,
      "author_name": "maryammohammadlou",
      "author_url": "",
      "post_date": "01/20/2025 09:54:31",
      "content": "<p>Today I joined the competition,  so probably I use the most updated train dataset. Still issue with coordination. train image 0002 has two polygons both Non-mitosis one of them in rage 512*512 but the second polygon far beyond </p>",
      "votes": null,
      "replies": [
        {
          "id": 3101063,
          "author_name": "maryammohammadlou",
          "author_url": "",
          "post_date": "01/20/2025 09:57:17",
          "content": "<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F16163090%2Fb2b5feacebdcee5cf0c73905fa7761ac%2FScreenshot%202025-01-20%20115700.png?generation=1737367032272456&amp;alt=media\" alt=\"\"></p>",
          "votes": null,
          "replies": [
            {
              "id": 3105905,
              "author_name": "seniai",
              "author_url": "",
              "post_date": "01/24/2025 05:43:41",
              "content": "<p>I've tried to decode the image_data in the JSON of training0002. It’s a 1015x1015 image, so I think this JSON provides the wrong value for imageHeight</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    },
    {
      "id": 3194818,
      "author_name": "casherjoel",
      "author_url": "",
      "post_date": "05/06/2025 10:35:21",
      "content": "<p>Hi!<br>\nI have trained some models and I'm having problems regarding the score of the test set. My models get satisfactory results on train and validation sets, but when predicting the test set, the performance is the same as a random one. I also created a test set from the train data to check for overfitting, and the results were similar to validation, so I don’t think it’s overfitting. Has anyone had this issue? Does anyone know why this might be happening?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3073874": "The image size of some data in the training data is inconsistent with the image size in the json annotation.",
    "3074662": "Thanks for pointing out Junqiang, we are tapping into the problem. Meanwhile, can you share the problematic data sample names you found?",
    "3074747": "the train dataset,i find imagesize is 512x512,but the related json file is 1015x1015\n(1015, 1015) (512, 512)\ntraining0002.json\n(1015, 1015) (512, 512)\ntraining0003.json\n(1015, 1015) (512, 512)\ntraining0004.json\n(1015, 1015) (512, 512)\ntraining0005.json\n(1015, 1015) (512, 512)\ntraining0006.json\n(1015, 1015) (512, 512)\ntraining0007.json\n(1015, 1015) (512, 512)\ntraining0008.json\n(1015, 1015) (512, 512)\ntraining0009.json\n(1015, 1015) (512, 512)\ntraining0010.json\n(1015, 1015) (512, 512)\ntraining0011.json\n(1015, 1015) (512, 512)\ntraining0012.json\n(1015, 1015) (512, 512)\ntraining0013.json\n(1015, 1015) (512, 512)\ntraining0014.json\n(1015, 1015) (512, 512)\ntraining0015.json\n(1015, 1015) (512, 512)\ntraining0016.json\n(1015, 1015) (512, 512)\ntraining0017.json\n(1015, 1015) (512, 512)\ntraining0018.json\n(1015, 1015) (512, 512)\ntraining0019.json\n(1015, 1015) (512, 512)\ntraining0020.json\n(1015, 1015) (512, 512)\ntraining0021.json\n(1015, 1015) (512, 512)\ntraining0022.json\n(1015, 1015) (512, 512)\ntraining0023.json\n(1015, 1015) (512, 512)\ntraining0024.json\n(1015, 1015) (512, 512)\ntraining0025.json\n(1015, 1015) (512, 512)\ntraining0026.json\n(1015, 1015) (512, 512)\ntraining0027.json\n(1015, 1015) (512, 512)\ntraining0028.json\n(1015, 1015) (512, 512)\ntraining0029.json\n(1015, 1015) (512, 512)\ntraining0030.json\n(1015, 1015) (512, 512)\ntraining0031.json\n(1015, 1015) (512, 512)\ntraining0032.json\n(1015, 1015) (512, 512)\ntraining0033.json\n(1015, 1015) (512, 512)\ntraining0034.json\n(1015, 1015) (512, 512)\ntraining0035.json\n(1015, 1015) (512, 512)\ntraining0036.json\n(1015, 1015) (512, 512)\ntraining0037.json\n(1015, 1015) (512, 512)\ntraining0038.json\n(1015, 1015) (512, 512)\ntraining0039.json\n(1015, 1015) (512, 512)\ntraining0040.json\n(1015, 1015) (512, 512)\ntraining0041.json\n(1015, 1015) (512, 512)\ntraining0042.json\n(1015, 1015) (512, 512)\ntraining0043.json\n(1015, 1015) (512, 512)\ntraining0044.json\n(1015, 1015) (512, 512)\ntraining0045.json\n(1015, 1015) (512, 512)\ntraining0046.json\n(1015, 1015) (512, 512)\ntraining0047.json\n(1015, 1015) (512, 512)\ntraining0048.json\n(1015, 1015) (512, 512)\ntraining0049.json\n(1015, 1015) (512, 512)\ntraining0050.json\n(1015, 1015) (512, 512)\ntraining0051.json\n(1015, 1015) (512, 512)\ntraining0052.json\n(1015, 1015) (512, 512)\ntraining0053.json\n(1015, 1015) (512, 512)\ntraining0054.json\n(1015, 1015) (512, 512)\ntraining0055.json\n(1015, 1015) (512, 512)\ntraining0056.json\n(1015, 1015) (512, 512)\ntraining0057.json\n(1015, 1015) (512, 512)\ntraining0058.json\n(1015, 1015) (512, 512)\ntraining0059.json\n(1015, 1015) (512, 512)\ntraining0060.json\n(1015, 1015) (512, 512)\ntraining0061.json\n(1015, 1015) (512, 512)\ntraining0062.json\n(1015, 1015) (512, 512)\ntraining0063.json\n(1015, 1015) (512, 512)\ntraining0064.json\n(1015, 1015) (512, 512)\ntraining0065.json\n(1015, 1015) (512, 512)\ntraining0066.json\n(1015, 1015) (512, 512)\ntraining0067.json\n(1015, 1015) (512, 512)\ntraining0068.json\n(1015, 1015) (512, 512)\ntraining0069.json\n(1015, 1015) (512, 512)\ntraining0070.json\n(1015, 1015) (512, 512)\ntraining0071.json\n(1015, 1015) (512, 512)\ntraining0072.json\n(1015, 1015) (512, 512)\ntraining0073.json\n(1015, 1015) (512, 512)\ntraining0074.json\n(1015, 1015) (512, 512)\ntraining0075.json\n(1015, 1015) (512, 512)\ntraining0076.json\n(1015, 1015) (512, 512)\ntraining0077.json\n(1015, 1015) (512, 512)\ntraining0078.json\n(1015, 1015) (512, 512)\ntraining0079.json\n(1015, 1015) (512, 512)\ntraining0080.json\n(1015, 1015) (512, 512)\ntraining0081.json\n(1015, 1015) (512, 512)\ntraining0082.json\n(1015, 1015) (512, 512)\ntraining0083.json\n(1015, 1015) (512, 512)\ntraining0084.json\n(1015, 1015) (512, 512)\ntraining0085.json\n(1015, 1015) (512, 512)\ntraining0086.json\n(1015, 1015) (512, 512)\ntraining0087.json\n(1015, 1015) (512, 512)\ntraining0088.json\n(1015, 1015) (512, 512)\ntraining0089.json\n(1015, 1015) (512, 512)\ntraining0090.json\n(1015, 1015) (512, 512)\ntraining0091.json\n(1015, 1015) (512, 512)\ntraining0092.json\n(1015, 1015) (512, 512)\ntraining0093.json\n(1015, 1015) (512, 512)\ntraining0094.json\n(1015, 1015) (512, 512)\ntraining0095.json\n(1015, 1015) (512, 512)\ntraining0096.json\n(1015, 1015) (512, 512)\ntraining0097.json\n(1015, 1015) (512, 512)\ntraining0098.json\n(1015, 1015) (512, 512)\ntraining0099.json\n(1015, 1015) (512, 512)\ntraining0100.json\n(1015, 1015) (512, 512)\ntraining0101.json\n(1015, 1015) (512, 512)\ntraining0102.json\n(1015, 1015) (512, 512)\ntraining0103.json\n(1015, 1015) (512, 512)\ntraining0104.json\n(1015, 1015) (512, 512)\ntraining0105.json\n(1015, 1015) (512, 512)\ntraining0106.json\n(1015, 1015) (512, 512)\ntraining0107.json\n(1015, 1015) (512, 512)\ntraining0108.json\n(1015, 1015) (512, 512)\ntraining0109.json\n(1015, 1015) (512, 512)\ntraining0110.json\n(1015, 1015) (512, 512)\ntraining0111.json\n(1015, 1015) (512, 512)\ntraining0112.json\n(1015, 1015) (512, 512)\ntraining0113.json\n(1015, 1015) (512, 512)\ntraining0114.json\n(1015, 1015) (512, 512)\ntraining0115.json\n(1015, 1015) (512, 512)\ntraining0116.json\n(1015, 1015) (512, 512)\ntraining0117.json\n(1015, 1015) (512, 512)\ntraining0118.json\n(1015, 1015) (512, 512)\ntraining0119.json\n(1015, 1015) (512, 512)\ntraining0120.json\n(1015, 1015) (512, 512)\ntraining0121.json\n(1015, 1015) (512, 512)\ntraining0122.json\n(1015, 1015) (512, 512)\ntraining0123.json\n(1015, 1015) (512, 512)\ntraining0124.json\n(1015, 1015) (512, 512)\ntraining0125.json\n(1015, 1015) (512, 512)\ntraining0126.json\n(1015, 1015) (512, 512)\ntraining0127.json\n(1015, 1015) (512, 512)\ntraining0128.json\n(1015, 1015) (512, 512)\ntraining0129.json\n(1015, 1015) (512, 512)\ntraining0130.json\n(1015, 1015) (512, 512)\ntraining0131.json\n(1015, 1015) (512, 512)\ntraining0132.json\n(1015, 1015) (512, 512)\ntraining0133.json\n(1015, 1015) (512, 512)\ntraining0134.json\n(1015, 1015) (512, 512)\ntraining0135.json\n(1015, 1015) (512, 512)\ntraining0136.json\n(1015, 1015) (512, 512)\ntraining0137.json\n(1015, 1015) (512, 512)\ntraining0138.json\n(1015, 1015) (512, 512)\ntraining0139.json\n(1015, 1015) (512, 512)\ntraining0140.json\n(1015, 1015) (512, 512)\ntraining0141.json\n(1015, 1015) (512, 512)\ntraining0142.json\n(1015, 1015) (512, 512)\ntraining0143.json\n(1015, 1015) (512, 512)\ntraining0144.json\n(1015, 1015) (512, 512)\ntraining0145.json\n(1015, 1015) (512, 512)\ntraining0146.json\n(1015, 1015) (512, 512)\ntraining0147.json\n(1015, 1015) (512, 512)\ntraining0148.json\n(1015, 1015) (512, 512)\ntraining0149.json\n(1015, 1015) (512, 512)\ntraining0150.json\n(1015, 1015) (512, 512)\ntraining0151.json\n(1015, 1015) (512, 512)\ntraining0152.json\n(1015, 1015) (512, 512)\ntraining0153.json\n(1015, 1015) (512, 512)\ntraining0154.json\n(1015, 1015) (512, 512)\ntraining0155.json\n(1015, 1015) (512, 512)\ntraining0156.json\n(1015, 1015) (512, 512)\ntraining0157.json\n(1015, 1015) (512, 512)\ntraining0158.json\n(1015, 1015) (512, 512)\ntraining0159.json\n(1015, 1015) (512, 512)\ntraining0160.json\n(1015, 1015) (512, 512)\ntraining0161.json\n(1015, 1015) (512, 512)\ntraining0162.json\n(1015, 1015) (512, 512)\ntraining0163.json\n(1015, 1015) (512, 512)\ntraining0164.json\n(1015, 1015) (512, 512)\ntraining0165.json\n(1015, 1015) (512, 512)\ntraining0166.json\n(1015, 1015) (512, 512)\ntraining0167.json\n(1015, 1015) (512, 512)\ntraining0168.json\n(1015, 1015) (512, 512)\ntraining0169.json\n(1015, 1015) (512, 512)\ntraining0170.json\n(1015, 1015) (512, 512)\ntraining0171.json\n(1015, 1015) (512, 512)\ntraining0172.json\n(1015, 1015) (512, 512)\ntraining0173.json\n(1015, 1015) (512, 512)\ntraining0174.json\n(1015, 1015) (512, 512)\ntraining0175.json\n(1015, 1015) (512, 512)\ntraining0176.json\n(1015, 1015) (512, 512)\ntraining0177.json\n(1015, 1015) (512, 512)\ntraining0178.json\n(1015, 1015) (512, 512)\ntraining0179.json",
    "3074749": "the test dataset,i find imagesize is 512x512,but the related json file is 1015x1015\n(1015, 1015) (512, 512)\ntesting0002.json\n(1015, 1015) (512, 512)\ntesting0003.json\n(1015, 1015) (512, 512)\ntesting0004.json\n(1015, 1015) (512, 512)\ntesting0005.json\n(1015, 1015) (512, 512)\ntesting0006.json\n(1015, 1015) (512, 512)\ntesting0007.json\n(1015, 1015) (512, 512)\ntesting0008.json\n(1015, 1015) (512, 512)\ntesting0009.json\n(1015, 1015) (512, 512)\ntesting0010.json\n(1015, 1015) (512, 512)\ntesting0011.json\n(1015, 1015) (512, 512)\ntesting0012.json\n(1015, 1015) (512, 512)\ntesting0013.json\n(1015, 1015) (512, 512)\ntesting0014.json\n(1015, 1015) (512, 512)\ntesting0015.json\n(1015, 1015) (512, 512)\ntesting0016.json\n(1015, 1015) (512, 512)\ntesting0017.json\n(1015, 1015) (512, 512)\ntesting0018.json",
    "3075502": "Thanks for the time to find out the unmatchable files. We have come up with a solution to solve the unmatching problem. Please use this code below to *up-sample* the 512-by-512 images to 1015-by-1015 dimensions. Let me know if you need any further help.\n\n\n\n    import cv2\n    import os\n\n    # Set the input and output folder paths\n    input_dir = r\"input_images\"\n    output_dir = r\"output_images\"\n\n    # If the output folder does not exist, create it\n    if not os.path.exists(output_dir):\n        os.makedirs(output_dir)\n\n    # Target size (1015x1015)\n    target_size = (1015, 1015)\n\n    # Get all .jpg image files from the input folder\n    image_files = [f for f in os.listdir(input_dir) if f.lower().endswith('.jpg')]\n\n    # Batch upsampling\n    for img_name in image_files:\n        input_path = os.path.join(input_dir, img_name)\n        output_path = os.path.join(output_dir, img_name.replace('.jpg', '_resized.jpg'))\n\n        # Read the original image\n        img = cv2.imread(input_path)\n        \n        if img is None:\n            print(f\"Unable to read image: {input_path}, skipping...\")\n            continue\n        \n        # Use bicubic interpolation for upsampling\n        resized_img = cv2.resize(img, target_size, interpolation=cv2.INTER_CUBIC)\n        \n        # Save the processed image\n        cv2.imwrite(output_path, resized_img)\n        print(f\"Processing completed: {output_path}\")",
    "3075559": "thank you,it's working",
    "3081223": "I think there is some issue with the JSON  files. I was checking the annotations of file number \"training0016.json\" via labelme, but the image it shows is actually \"training0015.jpg\". To verify, I double-checked it by plotting it using Python. I am assuming the numbering of JSON is wrong.",
    "3082027": "Sorry for the repeated problems with this dataset, and thank you for pointing out the unmatch issue. After discussion, we will redo the annotation on the 1015-by-1015 samples; please give us two to three days to finish. While waiting for the updated version, you can still train your model based on 512-by-512 samples, those are not going to be re-annotated. Thank you again, and don't hesitate to get in touch with us if you have any further questions. Happy holiday ;-)",
    "3082937": "Thanks all for your engagement.\n\nWe have updated our dataset to version **Data_122824**. Please use the most up-to-date version to train your model.\nWhat has been changed in this version:\n\n1. Re-annotated unclear images for better training results.\n2. Fixed the unmatched JSON and JPG files\n3. Added one more test sample (you may need to use the most up-to-date Submission_template.csv)\n\n**The data format and evaluation metric remained the same. You don't have to worry about changing your model structure.**",
    "3088132": "I'm using Data_122824 but yet even training0001.jpg seems to have a problem. The image size is 512x512 but JSON coordinates include larger than 512, indicating it's actually 1015.",
    "3088687": "Thanks, we are looking into it.",
    "3090770": "Thanks for waiting, after discussion with our team, please use the code in this question to preprocess the dimension.",
    "3092394": "excuseme, I find the wbsite data explorer are correct shape ,but not equals the website when i download , for example training0003.json in Data_122824 folder",
    "3096317": "Hi,\n\nI also encountered an issue with unmatched JSON and JPG files in the latest updated dataset, Data_122824.\n\nFor example, the image \"training0010.jpg\" in the left side of Fig. 1 is the original version, but after loading it into LabelMe, it displays a different case. This discrepancy may lead to annotation errors when visualizing the labels from the corresponding JSON files in the original image (as shown in Fig. 2). \n\nI’ve found that this problem affects several other files, including \"training0010.jpg,\" \"training0080.jpg,\" \"training0099.jpg,\" \"training0102.jpg,\" \"training0114.jpg,\" and others.\n\nHas anyone else encountered and solved this problem? Or could you advise on how I can fix the JSON files?\n\nThank you for your assistance.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4596430%2Fc6288be968947b9d066e23860fa68611%2F1.png?generation=1736844271316451&alt=media)\nfig1. the original image and showed in lableme\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F4596430%2F7cdb7f2cf4de06b95f77eaedeb4c780f%2F2.png?generation=1736844283864866&alt=media)  \nfig2. the visualized image (case: training0010.jpg)",
    "3097256": "Dear @wentaolimda Is there any solution i missed or you are still progress this issue? Thank you very much",
    "3097713": "Sorry for the late reply as our team has been working continuously on this problem. Right now, we don't know why exactly the lableme shows different images (actually it always pointed to training0001 from the problematic ones). While we are still working on the problem, please ignore these images for now.",
    "3098078": "Thank you for your reply. I understand now.",
    "3101061": "Today I joined the competition,  so probably I use the most updated train dataset. Still issue with coordination. train image 0002 has two polygons both Non-mitosis one of them in rage 512*512 but the second polygon far beyond",
    "3101063": "![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F16163090%2Fb2b5feacebdcee5cf0c73905fa7761ac%2FScreenshot%202025-01-20%20115700.png?generation=1737367032272456&alt=media)",
    "3105905": "I've tried to decode the image_data in the JSON of training0002. It’s a 1015x1015 image, so I think this JSON provides the wrong value for imageHeight",
    "3194818": "Hi!\nI have trained some models and I'm having problems regarding the score of the test set. My models get satisfactory results on train and validation sets, but when predicting the test set, the performance is the same as a random one. I also created a test set from the train data to check for overfitting, and the results were similar to validation, so I don’t think it’s overfitting. Has anyone had this issue? Does anyone know why this might be happening?"
  },
  "source": "meta"
}