{
  "total": 117,
  "topics": [
    {
      "id": 329650,
      "title": "Competition is Finalized - Congrats to our Winners, Recap",
      "comment_count": 2,
      "views": 0,
      "votes": 22
    },
    {
      "id": 328854,
      "title": "IMC2022 is over! Recap + Workshop prep",
      "comment_count": 6,
      "views": 0,
      "votes": 29
    },
    {
      "id": 316925,
      "title": "Welcome to the 2022 Image Matching Challenge!",
      "comment_count": 20,
      "views": 0,
      "votes": 50
    },
    {
      "id": 316922,
      "title": "Looking for a Team Megathread",
      "comment_count": 56,
      "views": 0,
      "votes": 23
    },
    {
      "id": 316921,
      "title": "New to Kaggle or Machine Learning? Come Say Hi!",
      "comment_count": 6,
      "views": 0,
      "votes": 30
    },
    {
      "id": 579086,
      "title": "Basic Idea of LoFTR",
      "comment_count": 0,
      "views": 0,
      "votes": -1
    },
    {
      "id": 329131,
      "title": "1st Place Solution",
      "comment_count": 39,
      "views": 0,
      "votes": 116
    },
    {
      "id": 316975,
      "title": "Intro to 3D Image Matching, what is a Fundamental Matrix?",
      "comment_count": 22,
      "views": 0,
      "votes": 94
    },
    {
      "id": 401124,
      "title": "Image Matching Challenge 2023",
      "comment_count": 0,
      "views": 0,
      "votes": 2
    },
    {
      "id": 393241,
      "title": "problem with using the ground truth ",
      "comment_count": 0,
      "views": 0,
      "votes": 0
    },
    {
      "id": 329540,
      "title": "3rd Place Solution 😎",
      "comment_count": 2,
      "views": 0,
      "votes": 19
    },
    {
      "id": 329317,
      "title": "2nd Place Solution, ensemble of old and new methods",
      "comment_count": 15,
      "views": 0,
      "votes": 45
    },
    {
      "id": 328805,
      "title": "4th Public LB 0.852 -> 6th Private LB 0.850",
      "comment_count": 21,
      "views": 0,
      "votes": 53
    },
    {
      "id": 328976,
      "title": "Are the results of this competition aligned with initial goals of organizers?",
      "comment_count": 4,
      "views": 0,
      "votes": 16
    },
    {
      "id": 329566,
      "title": "14th Place Solution",
      "comment_count": 0,
      "views": 0,
      "votes": 12
    },
    {
      "id": 329015,
      "title": "7th place solution (brief summary)",
      "comment_count": 9,
      "views": 0,
      "votes": 24
    },
    {
      "id": 328796,
      "title": "[9th place] Detailed report",
      "comment_count": 23,
      "views": 0,
      "votes": 51
    },
    {
      "id": 329002,
      "title": "23rd place solution",
      "comment_count": 8,
      "views": 0,
      "votes": 12
    },
    {
      "id": 328903,
      "title": "10th Place solution (brief summary)",
      "comment_count": 6,
      "views": 0,
      "votes": 35
    },
    {
      "id": 328999,
      "title": "Trust CV: An effective use of prepared train data",
      "comment_count": 1,
      "views": 0,
      "votes": 18
    },
    {
      "id": 328798,
      "title": "4th place solution (brief summary)",
      "comment_count": 10,
      "views": 0,
      "votes": 43
    },
    {
      "id": 328844,
      "title": "30th Place 827 Private LB with image size 840",
      "comment_count": 2,
      "views": 0,
      "votes": 15
    },
    {
      "id": 328848,
      "title": "26th place",
      "comment_count": 6,
      "views": 0,
      "votes": 14
    },
    {
      "id": 329279,
      "title": "Why no one fine tuned a model?",
      "comment_count": 4,
      "views": 0,
      "votes": 7
    },
    {
      "id": 328803,
      "title": "17th place solution - Upscaling & Models",
      "comment_count": 10,
      "views": 0,
      "votes": 29
    },
    {
      "id": 328896,
      "title": "21st place solution summary",
      "comment_count": 5,
      "views": 0,
      "votes": 23
    },
    {
      "id": 328820,
      "title": "VSAC - anyone want to use on Kagle?",
      "comment_count": 11,
      "views": 0,
      "votes": 27
    },
    {
      "id": 328817,
      "title": "[28th] Baby Kagglers Deus Lab",
      "comment_count": 8,
      "views": 0,
      "votes": 25
    },
    {
      "id": 328887,
      "title": "11th Place (Public: 0.847 / Private: 0.845)",
      "comment_count": 3,
      "views": 0,
      "votes": 20
    },
    {
      "id": 328982,
      "title": "18th place - Ensemble of 3 matchers + bonus crazy ideas that didn't work",
      "comment_count": 0,
      "views": 0,
      "votes": 20
    },
    {
      "id": 328865,
      "title": "A useful augmentation",
      "comment_count": 4,
      "views": 0,
      "votes": 17
    },
    {
      "id": 328864,
      "title": "Have any teams improved their scores by using the training data?",
      "comment_count": 10,
      "views": 0,
      "votes": 9
    },
    {
      "id": 329012,
      "title": "37th place solution",
      "comment_count": 0,
      "views": 0,
      "votes": 13
    },
    {
      "id": 329157,
      "title": "Experiments with SegFormer for Semantic Segmentation.",
      "comment_count": 0,
      "views": 0,
      "votes": 7
    },
    {
      "id": 328861,
      "title": "What's your best single model? 0.817 in private LB",
      "comment_count": 5,
      "views": 0,
      "votes": 11
    },
    {
      "id": 328653,
      "title": "Final day! Good luck everyone with the final work/submissions 😊",
      "comment_count": 3,
      "views": 0,
      "votes": 23
    },
    {
      "id": 327755,
      "title": "Sudden 30+ positions leaps on top of the LB are... interesting",
      "comment_count": 36,
      "views": 0,
      "votes": 43
    },
    {
      "id": 328868,
      "title": "[77th Place] Solo, LOFTR - SE - Matchformer / Thoughts on F calculation",
      "comment_count": 11,
      "views": 0,
      "votes": 10
    },
    {
      "id": 328888,
      "title": "27th place 0.823 public lb - 0.828 private lb",
      "comment_count": 0,
      "views": 0,
      "votes": 12
    },
    {
      "id": 328877,
      "title": "On- & offline installation of VSAC, MAGSAC++.....",
      "comment_count": 0,
      "views": 0,
      "votes": 8
    },
    {
      "id": 328715,
      "title": "Don't forget to share your keypoints:)",
      "comment_count": 0,
      "views": 0,
      "votes": 9
    },
    {
      "id": 328765,
      "title": "Is there any way to select the final submission notebooks not being the team leader?",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 328376,
      "title": "Extended deadline?",
      "comment_count": 9,
      "views": 0,
      "votes": 4
    },
    {
      "id": 328626,
      "title": "Leakage of Public/Private test batch_id?",
      "comment_count": 32,
      "views": 0,
      "votes": -5
    },
    {
      "id": 328317,
      "title": "why my notebook doesn't show image and code?",
      "comment_count": 6,
      "views": 0,
      "votes": 7
    },
    {
      "id": 328732,
      "title": "Can I select a notebook that's still running for final submission",
      "comment_count": 3,
      "views": 0,
      "votes": 2
    },
    {
      "id": 328216,
      "title": "Is it possible that my method works fine on the test data of public leaderboard, but timeout or out of memory on private leaderboard?",
      "comment_count": 2,
      "views": 0,
      "votes": 5
    },
    {
      "id": 328307,
      "title": "How long to run inference per image for submission?",
      "comment_count": 2,
      "views": 0,
      "votes": 5
    },
    {
      "id": 328630,
      "title": "Could you please provide the maximum size of the images in the test data?",
      "comment_count": 1,
      "views": 0,
      "votes": 2
    },
    {
      "id": 327568,
      "title": "What's the best way to computet F without droping outliers? ",
      "comment_count": 3,
      "views": 0,
      "votes": 8
    },
    {
      "id": 328874,
      "title": "CM calculation",
      "comment_count": 0,
      "views": 0,
      "votes": -2
    },
    {
      "id": 326110,
      "title": "Notebook Threw Exception debugging?",
      "comment_count": 37,
      "views": 0,
      "votes": 10
    },
    {
      "id": 327504,
      "title": "Sharing my project structure for Kaggle competitions",
      "comment_count": 3,
      "views": 0,
      "votes": 18
    },
    {
      "id": 326700,
      "title": "SOTA Robust Estimators in Motion Segmentation",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 327320,
      "title": "Is the first place winner going to speak at CVPR?",
      "comment_count": 2,
      "views": 0,
      "votes": 9
    },
    {
      "id": 327113,
      "title": "How random is your score?",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 326621,
      "title": "Are password protected datasets allowed?",
      "comment_count": 3,
      "views": 0,
      "votes": 6
    },
    {
      "id": 326491,
      "title": "What is needed to train/finetune kornia LoFTr (running into errors)?",
      "comment_count": 1,
      "views": 0,
      "votes": 6
    },
    {
      "id": 325531,
      "title": "Information about the test set",
      "comment_count": 6,
      "views": 0,
      "votes": 14
    },
    {
      "id": 326396,
      "title": "segmentation immovable object by DeepLab",
      "comment_count": 3,
      "views": 0,
      "votes": 5
    },
    {
      "id": 320219,
      "title": "Finetune LoFtr over competition dataset",
      "comment_count": 36,
      "views": 0,
      "votes": 70
    },
    {
      "id": 321360,
      "title": "OpenGlue - Open Source Pipeline for Image Matching",
      "comment_count": 8,
      "views": 0,
      "votes": 20
    },
    {
      "id": 325955,
      "title": "An easy way saving GPU time when submitting",
      "comment_count": 0,
      "views": 0,
      "votes": 12
    },
    {
      "id": 326214,
      "title": "Submission Scoring Error",
      "comment_count": 0,
      "views": 0,
      "votes": 3
    },
    {
      "id": 326638,
      "title": "Overviewの日本語訳（Japanese translation）",
      "comment_count": 0,
      "views": 0,
      "votes": -12
    },
    {
      "id": 323364,
      "title": "Competition Rules: Use images other than pair?",
      "comment_count": 6,
      "views": 0,
      "votes": 11
    },
    {
      "id": 321177,
      "title": "Code and pretrained models for SOTA matcher released!",
      "comment_count": 13,
      "views": 0,
      "votes": 26
    },
    {
      "id": 325118,
      "title": "Unstable score when inference",
      "comment_count": 3,
      "views": 0,
      "votes": 9
    },
    {
      "id": 325409,
      "title": "how can we get pose from a wrong epipolar assumption",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 325198,
      "title": "Question about LoFTR model fine-tuning?",
      "comment_count": 3,
      "views": 0,
      "votes": 7
    },
    {
      "id": 325374,
      "title": "Competition Objective",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 325405,
      "title": "The question about the submitting the .csv file",
      "comment_count": 4,
      "views": 0,
      "votes": 4
    },
    {
      "id": 322450,
      "title": "Line matching approaches",
      "comment_count": 11,
      "views": 0,
      "votes": 20
    },
    {
      "id": 324805,
      "title": "The World of RANSAC v2",
      "comment_count": 1,
      "views": 0,
      "votes": 16
    },
    {
      "id": 318143,
      "title": "Experiments with pretrained models",
      "comment_count": 39,
      "views": 0,
      "votes": 51
    },
    {
      "id": 318265,
      "title": "TFMatch: Local features in TensorFlow",
      "comment_count": 6,
      "views": 0,
      "votes": 19
    },
    {
      "id": 323403,
      "title": "Slightly off topic",
      "comment_count": 17,
      "views": 0,
      "votes": 15
    },
    {
      "id": 322093,
      "title": "How long does it take to get the score for your submission",
      "comment_count": 7,
      "views": 0,
      "votes": 12
    },
    {
      "id": 324370,
      "title": "How accurate are the covisibility values in the training dataset?",
      "comment_count": 1,
      "views": 0,
      "votes": 5
    },
    {
      "id": 324410,
      "title": "Traditional Methods or Deep Learning?",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 318022,
      "title": "The way to use training images",
      "comment_count": 7,
      "views": 0,
      "votes": 13
    },
    {
      "id": 317969,
      "title": "Additional dataset for this competition",
      "comment_count": 5,
      "views": 0,
      "votes": 37
    },
    {
      "id": 318577,
      "title": "LoFTR validation and submission score",
      "comment_count": 16,
      "views": 0,
      "votes": 25
    },
    {
      "id": 319609,
      "title": "How do we submite the .csv file?",
      "comment_count": 3,
      "views": 0,
      "votes": 5
    },
    {
      "id": 322489,
      "title": "Benchmark score for this IMC 2022 ",
      "comment_count": 5,
      "views": 0,
      "votes": 11
    },
    {
      "id": 322602,
      "title": "Is there a better normalization method?",
      "comment_count": 3,
      "views": 0,
      "votes": 7
    },
    {
      "id": 320868,
      "title": "Winning solutions of past kaggle challenges involving Retrieval & Recognition",
      "comment_count": 1,
      "views": 0,
      "votes": 11
    },
    {
      "id": 318267,
      "title": "Correspondence filtering methods & learned matching",
      "comment_count": 3,
      "views": 0,
      "votes": 26
    },
    {
      "id": 322609,
      "title": "Help needed - Unable to submit prediction file",
      "comment_count": 3,
      "views": 0,
      "votes": 3
    },
    {
      "id": 321758,
      "title": "How many pairs to use for validation?",
      "comment_count": 7,
      "views": 0,
      "votes": 13
    },
    {
      "id": 317597,
      "title": "Notebook Threw Exception",
      "comment_count": 7,
      "views": 0,
      "votes": 7
    },
    {
      "id": 320463,
      "title": "Segmentation train dataset.",
      "comment_count": 2,
      "views": 0,
      "votes": 8
    },
    {
      "id": 317051,
      "title": "🏆 Recent Related Competition Winning Solution 🏆",
      "comment_count": 4,
      "views": 0,
      "votes": 19
    },
    {
      "id": 317531,
      "title": "Training a model using personal compute",
      "comment_count": 2,
      "views": 0,
      "votes": 6
    },
    {
      "id": 317076,
      "title": " 3D Reconstruction🎉",
      "comment_count": 1,
      "views": 0,
      "votes": 12
    },
    {
      "id": 321310,
      "title": "Back to the Feature: Learning Robust Camera Localization from Pixels to Pose",
      "comment_count": 0,
      "views": 0,
      "votes": 7
    },
    {
      "id": 317292,
      "title": "A demo showing epipolar correspondences from a Fundamental matrix",
      "comment_count": 2,
      "views": 0,
      "votes": 35
    },
    {
      "id": 319577,
      "title": "difference in image format",
      "comment_count": 5,
      "views": 0,
      "votes": 4
    },
    {
      "id": 318319,
      "title": "Youtube Playlist to Understand the Competition",
      "comment_count": 4,
      "views": 0,
      "votes": 58
    },
    {
      "id": 320563,
      "title": "How to set Loss Function?",
      "comment_count": 1,
      "views": 0,
      "votes": 4
    },
    {
      "id": 320088,
      "title": "End-to-end deep learning approach?",
      "comment_count": 2,
      "views": 0,
      "votes": 8
    },
    {
      "id": 320260,
      "title": "Best Image Size",
      "comment_count": 0,
      "views": 0,
      "votes": 4
    },
    {
      "id": 318590,
      "title": "A Concern about the Evaluation Metric",
      "comment_count": 1,
      "views": 0,
      "votes": 22
    },
    {
      "id": 318209,
      "title": "(a few) Starter Resources with TF.Keras.",
      "comment_count": 7,
      "views": 0,
      "votes": 23
    },
    {
      "id": 318138,
      "title": "Test Set Pairs Where Covisibility Is ~0.0",
      "comment_count": 1,
      "views": 0,
      "votes": 10
    },
    {
      "id": 318126,
      "title": "I updated the \"EfficientNet\" datasets for this (and future) competitions",
      "comment_count": 0,
      "views": 0,
      "votes": 13
    },
    {
      "id": 317816,
      "title": "Would be this a Torch Competition? With Kornia and 3D models.",
      "comment_count": 4,
      "views": 0,
      "votes": 7
    },
    {
      "id": 318056,
      "title": "Loading pretrained LoFTR",
      "comment_count": 6,
      "views": 0,
      "votes": 5
    },
    {
      "id": 317302,
      "title": "IMC 2021 winning solutions papers",
      "comment_count": 1,
      "views": 0,
      "votes": 31
    },
    {
      "id": 317717,
      "title": "The Fundamental Matrix Song",
      "comment_count": 0,
      "views": 0,
      "votes": 18
    },
    {
      "id": 317372,
      "title": "Questions about previous challenge dataset",
      "comment_count": 3,
      "views": 0,
      "votes": 10
    },
    {
      "id": 317433,
      "title": "Minor correction in the data description",
      "comment_count": 1,
      "views": 0,
      "votes": 8
    },
    {
      "id": 317058,
      "title": "Some resources for starting this competition",
      "comment_count": 1,
      "views": 0,
      "votes": 26
    },
    {
      "id": 316960,
      "title": "Detector-Free Local Feature Matching w Transformer - first experiment",
      "comment_count": 0,
      "views": 0,
      "votes": 33
    },
    {
      "id": 317176,
      "title": "RANSAC benchmark collection",
      "comment_count": 0,
      "views": 0,
      "votes": 10
    },
    {
      "id": 317039,
      "title": "Notebook Exceeded Allowed Compute, Now What?",
      "comment_count": 18,
      "views": 0,
      "votes": 7
    },
    {
      "id": 317046,
      "title": "3D Landmarks & 3D Position & Recgnition [awesome paper]",
      "comment_count": 0,
      "views": 0,
      "votes": 9
    }
  ],
  "errors": []
}