This task targets at the image-based multi-pose virtual try-on. Specifically, given an input person image, a desired clothing image, and a desired pose, the participator is asked to design algorithm to transfer the desired clothing to the person image and manipulate human poses.
MPV (Multi-Pose Virtual try on) dataset, which consists of 35,687/13,524 person/clothes images, with the resolution of 256x192. Each person has different poses. We split them into the train/test set 52,236/10,544 three-tuples, respectively.
You can download the dataset at MPV(Google Drive).
For image-based multi-pose virtual try on, we use two metrics for evaluation:
It is worth to note that, the evaluation metric for the Development phase is SSIM while the evaluation metric for the Final phase is AMT.
Specifically, during the Development phase, We measure the SSIM score between the generated images and ground truth images. The person of each pair is wearing same clothes. The due date for the Development phase is May 1, 2020. The TOP 10 users will be invited to attend the Final phase.
During the Final phase, We conduct A/B tests deployed on AMT platform for user study. The person of each pair is wearing different clothes. We will pick the TOP 10 users of Development phase to participate in the AMT evaluation. We will replace the instructions of the testset, the 10 users must upload the new results for AMT evaluation before May 20, 2020. We will announce the AMT scores in May 30, 2020.
A folder named mtryon_results.zip (Click to download a template file) contains your predicted images with .jpg format. The number of images should be the same of our Testing set. Make sure these and then package the folder with zip format. Submit your results.zip and wait to see your rank.
Haoye Dong, Xiaodan Liang, Bochao Wang, Hanjiang Lai, Jia Zhu, and Jian Yin. Towards multi-pose guided virtual try-on network. InICCV, 2019.
For more information, please concate us at firstname.lastname@example.org or email@example.com.
Start: Feb. 20, 2020, midnight
May 20, 2020, midnight
You must be logged in to participate in competitions.Sign In