{
  "id": 507943,
  "title": "Statement on the Challenge Rule Issue Regarding the Inclusion of Public Submissions in the Evaluation of Phase 1 Results",
  "url": "/competitions/cvpr-metafood-3d-food-reconstruction-challenge/discussion/507943",
  "author_name": "metafood-cvpr",
  "post_date": "2024-05-28T00:06:25.647000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>As stated in our initial rules:<br>\n\"Participants are required to submit 20 .obj mesh files corresponding to the 20 selected food items. These submissions will undergo evaluation based on the reconstruction's shape and size accuracy.\"</p>\n<p>There were no mentions in the rules about distinguishing between public and private scores. We designed the public and private scores to provide participants with a reference and feedback to help them achieve good performance with their algorithms on the 20 selected food items. This design aims to give positive feedback and acknowledgment to participants for their efforts in solving this challenging problem. It also allows them to see their progress and be rewarded with better scores. The challenge itself encourages the community to advance algorithm development in this field.</p>\n<p>Additionally, we implemented the rule: \"Teams will be disqualified if their submitted 3D models do not accurately represent the object class's shape.\" This aspect will be carefully examined by the committee to ensure no cheating occurs. Furthermore, Phase 2 will further ensure the quality of the models.</p>\n<p>In conclusion, we, as the MetaFood organizing committee, have decided to adhere to the original rule, where all 20 models will be evaluated for both shape and size accuracy. We believe this decision is in the best interest of the current food computing research community.</p>\n<p>Sincerely,<br>\nMetaFood Workshop Organizing Committee</p>",
  "messages": [
    {
      "id": 2840048,
      "postDate": "2024-05-28T00:06:25.647Z",
      "content": "<p>As stated in our initial rules:<br>\n\"Participants are required to submit 20 .obj mesh files corresponding to the 20 selected food items. These submissions will undergo evaluation based on the reconstruction's shape and size accuracy.\"</p>\n<p>There were no mentions in the rules about distinguishing between public and private scores. We designed the public and private scores to provide participants with a reference and feedback to help them achieve good performance with their algorithms on the 20 selected food items. This design aims to give positive feedback and acknowledgment to participants for their efforts in solving this challenging problem. It also allows them to see their progress and be rewarded with better scores. The challenge itself encourages the community to advance algorithm development in this field.</p>\n<p>Additionally, we implemented the rule: \"Teams will be disqualified if their submitted 3D models do not accurately represent the object class's shape.\" This aspect will be carefully examined by the committee to ensure no cheating occurs. Furthermore, Phase 2 will further ensure the quality of the models.</p>\n<p>In conclusion, we, as the MetaFood organizing committee, have decided to adhere to the original rule, where all 20 models will be evaluated for both shape and size accuracy. We believe this decision is in the best interest of the current food computing research community.</p>\n<p>Sincerely,<br>\nMetaFood Workshop Organizing Committee</p>",
      "rawMarkdown": "As stated in our initial rules:\n\"Participants are required to submit 20 .obj mesh files corresponding to the 20 selected food items. These submissions will undergo evaluation based on the reconstruction's shape and size accuracy.\"\n\nThere were no mentions in the rules about distinguishing between public and private scores. We designed the public and private scores to provide participants with a reference and feedback to help them achieve good performance with their algorithms on the 20 selected food items. This design aims to give positive feedback and acknowledgment to participants for their efforts in solving this challenging problem. It also allows them to see their progress and be rewarded with better scores. The challenge itself encourages the community to advance algorithm development in this field.\n\nAdditionally, we implemented the rule: \"Teams will be disqualified if their submitted 3D models do not accurately represent the object class's shape.\" This aspect will be carefully examined by the committee to ensure no cheating occurs. Furthermore, Phase 2 will further ensure the quality of the models.\n\nIn conclusion, we, as the MetaFood organizing committee, have decided to adhere to the original rule, where all 20 models will be evaluated for both shape and size accuracy. We believe this decision is in the best interest of the current food computing research community.\n\nSincerely,\nMetaFood Workshop Organizing Committee",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2840048": "As stated in our initial rules:\n\"Participants are required to submit 20 .obj mesh files corresponding to the 20 selected food items. These submissions will undergo evaluation based on the reconstruction's shape and size accuracy.\"\n\nThere were no mentions in the rules about distinguishing between public and private scores. We designed the public and private scores to provide participants with a reference and feedback to help them achieve good performance with their algorithms on the 20 selected food items. This design aims to give positive feedback and acknowledgment to participants for their efforts in solving this challenging problem. It also allows them to see their progress and be rewarded with better scores. The challenge itself encourages the community to advance algorithm development in this field.\n\nAdditionally, we implemented the rule: \"Teams will be disqualified if their submitted 3D models do not accurately represent the object class's shape.\" This aspect will be carefully examined by the committee to ensure no cheating occurs. Furthermore, Phase 2 will further ensure the quality of the models.\n\nIn conclusion, we, as the MetaFood organizing committee, have decided to adhere to the original rule, where all 20 models will be evaluated for both shape and size accuracy. We believe this decision is in the best interest of the current food computing research community.\n\nSincerely,\nMetaFood Workshop Organizing Committee"
  }
}