Competition

Evaluation Criteria

Track-1 | Completion

We evaluate the reconstruction accuracy by computing the Chamfer Distance between the predicted complete shape (P) and the ground truth shape (Q) as below. The results are averaged across the whole test set for overall evaluation criteria.

Track-2 | Registration

We evaluate the reconstruction accuracy by computing the differences between the predicted Transformation (T_pred) and the ground truth Transformation (T_gt) as below. The metric is defined as a weighted summation (denoted as MSE on the leaderboard) for 1) rotation angle differences and 2) translation differences. The results are averaged across the whole test set for overall evaluation criteria.

Please visit our gihub project for details.

Terms and Conditions

General Rules

The MVP Challenge 2021 will be around eight weeks. The challenge will start together with ICCV 2021, the 3rd Workshop on Sensing, Understanding and Synthesizing Humans. Participants are restricted to train their algorithms on the publicly available MVP training dataset. A hidden test set is used for online evaluation and for maintaining a public leaderboard. The final awards will be revealed around Oct. 2021.

When participating in the competition, please be reminded that:

• Results in the correct format must be uploaded to the evaluation server. The Evaluation page lists detailed information regarding how results will be evaluated.
• Each entry must be associated with a team and provide its affiliation.
• The online evaluation results must be submitted through this CodaLab competition site of the MVP Challenge. The participants can conduct five online evaluations per day. The organizer will offer some hints/codes to ensure a successful submission.
• Using multiple accounts to increase the number of submissions and private sharing outside teams are strictly prohibited.
• The organizer reserves the absolute right to disqualify entries that are incomplete or illegible, late entries, or entries that violate the rules.
• The organizer reserves the right to adjust the competition schedule and rules based on situations.
• The best entry of each team will be public on the leaderboard at all times.
• To compete for awards, the participants must submit a technical report to describe their methods, and they are encouraged to provide detailed ablation studies on the Validation set. There is no other publication requirement.

Terms of Use

Before downloading and using the MVP dataset, please agree to the following terms of use. You, your employer, and your affiliations are referred to as "User." The authors and their affiliations, SenseTime, are referred to as "Producer."

• The dataset is used for non-commercial/non-profit research purposes only.
• All the data in MVP dataset can be used for academic purposes. However, the Producer is NOT responsible for any further use in a defamatory, pornographic, or any other unlawful manner, or violation of any applicable regulations or laws.
• The User takes full responsibility for any consequence caused by his/her use of the dataset in any form and shall defend and indemnify the Producer against all claims arising from such services.
• The User should NOT distribute, copy, reproduce, disclose, assign, sublicense, embed, host, transfer, sell, trade, or resell any portion of the dataset to any third party for any purpose.
• This agreement is effective for any potential User of the MVP dataset upon the date that the User first accesses the dataset in any form.
• For using MVP dataset, please cite the following paper:

  @inproceedings{pan2021variational,
      title={Variational Relational Point Completion Network},
      author={Pan, Liang and Chen, Xinyi and Cai, Zhongang and Zhang, Junzhe and Zhao, Haiyu and Yi, Shuai and Liu, Ziwei},
      journal={arXiv preprint arXiv:2104.10154},
      year={2021}
    }