{
  "id": 613746,
  "title": "Question about baseline results and dataset usage",
  "url": "/competitions/2026-icassp-hyper-object-challenge-track-1/discussion/613746",
  "author_name": "",
  "post_date": "2025-10-29T12:23:23.586344400Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Dear organizers,</p>\n<p>I have a couple of questions regarding the Track 1 leaderboard and evaluation setup:</p>\n<p>Were the leaderboard baseline results (e.g., Baseline Model (MST++) – 0.726, (HRNet) – 0.560, (HSCNN+) – 0.350) obtained using the official Track 1 dataset (training + test set)?</p>\n<p>I noticed that the provided baseline code in the official repository (the Raw2HSI model) does not correspond to the leaderboard models (MST++ / HRNet / HSCNN+). Could you please confirm whether the leaderboard scores were produced using the same dataset, and whether you plan to release the corresponding baseline implementations?</p>\n<p>Additionally, the leaderboard description mentions evaluation metrics based on RMSE, SAM, PSNR, SSIM, and ERGAS, while the Kaggle Evaluation page defines the official ranking metric  using SAM, SID, ERGAS, PSNR, SSIM, and ΔE00. Could you please clarify which metric combination is currently used for the official ranking?</p>\n<p>Thank you very much for your time and clarification.</p>",
  "messages": [
    {
      "id": "3308433",
      "postDate": "10/29/2025 12:23:23",
      "content": "<p>Dear organizers,</p>\n<p>I have a couple of questions regarding the Track 1 leaderboard and evaluation setup:</p>\n<p>Were the leaderboard baseline results (e.g., Baseline Model (MST++) – 0.726, (HRNet) – 0.560, (HSCNN+) – 0.350) obtained using the official Track 1 dataset (training + test set)?</p>\n<p>I noticed that the provided baseline code in the official repository (the Raw2HSI model) does not correspond to the leaderboard models (MST++ / HRNet / HSCNN+). Could you please confirm whether the leaderboard scores were produced using the same dataset, and whether you plan to release the corresponding baseline implementations?</p>\n<p>Additionally, the leaderboard description mentions evaluation metrics based on RMSE, SAM, PSNR, SSIM, and ERGAS, while the Kaggle Evaluation page defines the official ranking metric  using SAM, SID, ERGAS, PSNR, SSIM, and ΔE00. Could you please clarify which metric combination is currently used for the official ranking?</p>\n<p>Thank you very much for your time and clarification.</p>",
      "rawMarkdown": "Dear organizers,\n\nI have a couple of questions regarding the Track 1 leaderboard and evaluation setup:\n\nWere the leaderboard baseline results (e.g., Baseline Model (MST++) – 0.726, (HRNet) – 0.560, (HSCNN+) – 0.350) obtained using the official Track 1 dataset (training + test set)?\n\nI noticed that the provided baseline code in the official repository (the Raw2HSI model) does not correspond to the leaderboard models (MST++ / HRNet / HSCNN+). Could you please confirm whether the leaderboard scores were produced using the same dataset, and whether you plan to release the corresponding baseline implementations?\n\nAdditionally, the leaderboard description mentions evaluation metrics based on RMSE, SAM, PSNR, SSIM, and ERGAS, while the Kaggle Evaluation page defines the official ranking metric  using SAM, SID, ERGAS, PSNR, SSIM, and ΔE00. Could you please clarify which metric combination is currently used for the official ranking?\n\nThank you very much for your time and clarification.",
      "votes": null
    },
    {
      "id": "3313088",
      "postDate": "11/08/2025 16:18:26",
      "content": "<ol>\n<li><p>I presumed that you refer to the leaderboard in the website. That leaderboard is just a temporary placeholder when the website was built, and we will update the final leaderboard with the hidden test data that will be shared with the registered team. Please register here: <a href=\"https://docs.google.com/forms/d/e/1FAIpQLSemy6azp0ZBrT2I5lWZln7bj5kxfiZvgIr4lVevUz2i7Z7iLQ/viewform\" target=\"_blank\">Registration Form</a></p></li>\n<li><p>We will use the ranking metric as in the Kaggle Evaluation page. Sorry for the discrepancy, we will update the information in the website. </p></li>\n</ol>",
      "rawMarkdown": "1. I presumed that you refer to the leaderboard in the website. That leaderboard is just a temporary placeholder when the website was built, and we will update the final leaderboard with the hidden test data that will be shared with the registered team. Please register here: [Registration Form](https://docs.google.com/forms/d/e/1FAIpQLSemy6azp0ZBrT2I5lWZln7bj5kxfiZvgIr4lVevUz2i7Z7iLQ/viewform)\n\n2. We will use the ranking metric as in the Kaggle Evaluation page. Sorry for the discrepancy, we will update the information in the website.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3313088,
      "author_name": "hyperobject",
      "author_url": "",
      "post_date": "11/08/2025 16:18:26",
      "content": "<ol>\n<li><p>I presumed that you refer to the leaderboard in the website. That leaderboard is just a temporary placeholder when the website was built, and we will update the final leaderboard with the hidden test data that will be shared with the registered team. Please register here: <a href=\"https://docs.google.com/forms/d/e/1FAIpQLSemy6azp0ZBrT2I5lWZln7bj5kxfiZvgIr4lVevUz2i7Z7iLQ/viewform\" target=\"_blank\">Registration Form</a></p></li>\n<li><p>We will use the ranking metric as in the Kaggle Evaluation page. Sorry for the discrepancy, we will update the information in the website. </p></li>\n</ol>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "3308433": "Dear organizers,\n\nI have a couple of questions regarding the Track 1 leaderboard and evaluation setup:\n\nWere the leaderboard baseline results (e.g., Baseline Model (MST++) – 0.726, (HRNet) – 0.560, (HSCNN+) – 0.350) obtained using the official Track 1 dataset (training + test set)?\n\nI noticed that the provided baseline code in the official repository (the Raw2HSI model) does not correspond to the leaderboard models (MST++ / HRNet / HSCNN+). Could you please confirm whether the leaderboard scores were produced using the same dataset, and whether you plan to release the corresponding baseline implementations?\n\nAdditionally, the leaderboard description mentions evaluation metrics based on RMSE, SAM, PSNR, SSIM, and ERGAS, while the Kaggle Evaluation page defines the official ranking metric  using SAM, SID, ERGAS, PSNR, SSIM, and ΔE00. Could you please clarify which metric combination is currently used for the official ranking?\n\nThank you very much for your time and clarification.",
    "3313088": "1. I presumed that you refer to the leaderboard in the website. That leaderboard is just a temporary placeholder when the website was built, and we will update the final leaderboard with the hidden test data that will be shared with the registered team. Please register here: [Registration Form](https://docs.google.com/forms/d/e/1FAIpQLSemy6azp0ZBrT2I5lWZln7bj5kxfiZvgIr4lVevUz2i7Z7iLQ/viewform)\n\n2. We will use the ranking metric as in the Kaggle Evaluation page. Sorry for the discrepancy, we will update the information in the website."
  },
  "source": "meta"
}