Object landmark discovery through unsupervised adaptation

Sanchez, Enrique; Tzimiropoulos, Georgios

Abstract

This paper proposes a method to ease the unsupervised learning of object landmark detectors. Similarly to previous methods, our approach is fully unsupervised in a sense that it does not require or make any use of annotated landmarks for the target object category. Contrary to previous works, we do however assume that a landmark detector, which has already learned a structured representation for a given object category in a fully supervised manner, is available. Under this setting, our main idea boils down to adapting the given pre-trained network to the target object categories in a fully unsupervised manner. To this end, our method uses the pre-trained network as a core which remains frozen and does not get updated during training, and learns, in an unsupervised manner, only a projection matrix to perform the adaptation to the target categories. By building upon an existing structured representation learned in a supervised manner, the optimization problem solved by our method is much more constrained with significantly less parameters to learn which seems to be important for the case of unsupervised learning. We show that our method surpasses fully unsupervised techniques trained from scratch as well as a strong baseline based on fine-tuning, and produces state-of-the-art results on several datasets. Code can be found at tiny.cc/GitHub-Unsupervised

Citation

Sanchez, E., & Tzimiropoulos, G. (2019). Object landmark discovery through unsupervised adaptation. Advances in Neural Information Processing Systems, 32,

• Journal Article Type: Conference Paper
• Conference Name: Thirty-third Conference on Neural Information Processing Systems
• Start Date: Dec 8, 2019
• End Date: Dec 14, 2019
• Acceptance Date: Sep 4, 2019
• Publication Date: 2019
• Deposit Date: Nov 1, 2019
• Publicly Available Date: Nov 5, 2019
• Publisher: Massachusetts Institute of Technology Press
• Volume: 32
• Book Title: Advances in Neural Information Processing Systems
• Public URL: https://nottingham-repository.worktribe.com/output/3009966
• Publisher URL: http://papers.nips.cc/paper/9505-object-landmark-discovery-through-unsupervised-adaptation
• Related Public URLs: https://nips.cc/