Pose Transfer using a Single Spatial Transformation

Mohsen Tabejamaat, Francois Bremond

Abstract: In this paper, we address the problem of pose transfer. The goal is to generate a source image in a new target pose. The pose is already provided by a set of spatial landmarks. The transfer function is directly estimated from the difference between the landmarks given in the new target pose and the landmarks of the source image. Existing methods perform the task using two specialized networks, one to move the patches of the source sample and the other one to generate the new patches that are not visible in the source image. Contrary to these strategies, we develop an end-to-end trainable neural network that learns to estimate both these visible and invisible parts using a simple warping module. In other words, we propose a flow estimation method that not only displaces the patches to their new locations but also generates new pixels that are not visible in the source image, all in an unsupervised manner without the need for a ground-truth flow map. In this way, moving the patches and introducing new parts are unified into a single network, ensuring that an overall solution is achieved for these two mutual tasks. Additionally, this method avoids the need for a human observer to determine a trade-off between the performance of the two separated networks, thus avoiding a cartoonish addition of the new parts to the visible areas. Extensive experiments demonstrate the superiority of our method over state-of-the-art algorithms. We conduct our experiments on two well-known datasets: Deepfashion and Market1501.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Generative models