{
  "id": 483413,
  "title": "Uh-oh...my fru- fruit is gone",
  "url": "/competitions/csc532-fruits-categorical-classification-2024/discussion/483413",
  "author_name": "",
  "post_date": "2024-03-12T10:17:10.828579800Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Based on my experiment with the background remover using a segmentation model on the validation images, I found out that it also removes the fruits (the objects that we want to classify) in some cases. So, I think I should be more careful with my background removal method because my model is configured to classify fruits, not HANDS! I have an example below. </p>\n<h1>'DLAI5_Validation_fruits (12).jpg' before the background remover</h1>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5080265%2Fc77ec5d533e1c60b53d94d8bc6d3e0bf%2FDLAI5_Validation_fruits%20(12).jpg?generation=1710237493867767&amp;alt=media\"></p>\n<h1>'DLAI5_Validation_fruits (12).jpg' after the background remover.</h1>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5080265%2Fb531564809aba224aaf9b3b6fa0e78cf%2FDLAI5_Validation_fruits%20(12).jpg?generation=1710237370910161&amp;alt=media\"></p>",
  "messages": [
    {
      "id": "2693278",
      "postDate": "03/12/2024 10:17:10",
      "content": "<p>Based on my experiment with the background remover using a segmentation model on the validation images, I found out that it also removes the fruits (the objects that we want to classify) in some cases. So, I think I should be more careful with my background removal method because my model is configured to classify fruits, not HANDS! I have an example below. </p>\n<h1>'DLAI5_Validation_fruits (12).jpg' before the background remover</h1>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5080265%2Fc77ec5d533e1c60b53d94d8bc6d3e0bf%2FDLAI5_Validation_fruits%20(12).jpg?generation=1710237493867767&amp;alt=media\"></p>\n<h1>'DLAI5_Validation_fruits (12).jpg' after the background remover.</h1>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5080265%2Fb531564809aba224aaf9b3b6fa0e78cf%2FDLAI5_Validation_fruits%20(12).jpg?generation=1710237370910161&amp;alt=media\"></p>",
      "rawMarkdown": "Based on my experiment with the background remover using a segmentation model on the validation images, I found out that it also removes the fruits (the objects that we want to classify) in some cases. So, I think I should be more careful with my background removal method because my model is configured to classify fruits, not HANDS! I have an example below. \n\n# 'DLAI5_Validation_fruits (12).jpg' before the background remover\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5080265%2Fc77ec5d533e1c60b53d94d8bc6d3e0bf%2FDLAI5_Validation_fruits%20(12).jpg?generation=1710237493867767&alt=media)\n\n# 'DLAI5_Validation_fruits (12).jpg' after the background remover.\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5080265%2Fb531564809aba224aaf9b3b6fa0e78cf%2FDLAI5_Validation_fruits%20(12).jpg?generation=1710237370910161&alt=media)",
      "votes": null
    },
    {
      "id": "2693664",
      "postDate": "03/12/2024 15:26:31",
      "content": "<p><a href=\"https://www.kaggle.com/kaungmyatkyaw\" target=\"_blank\">@kaungmyatkyaw</a> : Nice observation!</p>",
      "rawMarkdown": "kaungmyatkyaw : Nice observation!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2693664,
      "author_name": "jonathanchan",
      "author_url": "",
      "post_date": "03/12/2024 15:26:31",
      "content": "<p><a href=\"https://www.kaggle.com/kaungmyatkyaw\" target=\"_blank\">@kaungmyatkyaw</a> : Nice observation!</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2693278": "Based on my experiment with the background remover using a segmentation model on the validation images, I found out that it also removes the fruits (the objects that we want to classify) in some cases. So, I think I should be more careful with my background removal method because my model is configured to classify fruits, not HANDS! I have an example below. \n\n# 'DLAI5_Validation_fruits (12).jpg' before the background remover\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5080265%2Fc77ec5d533e1c60b53d94d8bc6d3e0bf%2FDLAI5_Validation_fruits%20(12).jpg?generation=1710237493867767&alt=media)\n\n# 'DLAI5_Validation_fruits (12).jpg' after the background remover.\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F5080265%2Fb531564809aba224aaf9b3b6fa0e78cf%2FDLAI5_Validation_fruits%20(12).jpg?generation=1710237370910161&alt=media)",
    "2693664": "kaungmyatkyaw : Nice observation!"
  },
  "source": "meta"
}