This paper addresses the problem of unsupervised discovery of object landmarks. We take a different path compared to existing works, based on 2 novel perspectives: (1) Self-training: starting from generic keypoints, we propose a self-training approach where the goal is to learn a detector that improves itself, becoming more and more tuned to object landmarks. (2) Correspondence: we identify correspondence as a key objective for unsupervised landmark discovery and propose an optimization scheme which alternates between recovering object landmark correspondence across different images via clustering and learning an object landmark descriptor without labels. Compared to previous works, our approach can learn landmarks that are more flexible in terms of capturing large changes in viewpoint. We show the favourable properties of our method on a variety of difficult datasets including LS3D, BBCPose and Human3.6M. Code is available at https://github.com/malldimi1/UnsupervisedLandmarks.
Dimitrios Mallis (Computer Vision Laboratory - University of Nottingham)
Enrique Sanchez (Samsung AI Centre)
I am a Senior Research Scientist at Samsung AI Cambridge, UK. Prior to that, I was a Research Fellow at the University of Nottingham, from 2016 to 2019. I received my PhD degree in Computer Science from the University of Nottingham in 2017. I received my MEng in Telecommunication Engineering and MSc in Signal Theory and Communications from the University of Vigo (Spain), in 2009 and 2011, respectively.
Matthew Bell (University of Nottingham)
Georgios Tzimiropoulos (Queen Mary University of London)
More from the Same Authors
2021 Poster: Space-time Mixing Attention for Video Transformer »
Adrian Bulat · Juan Manuel Perez Rua · Swathikiran Sudhakaran · Brais Martinez · Georgios Tzimiropoulos
2019 Poster: Object landmark discovery through unsupervised adaptation »
Enrique Sanchez · Georgios Tzimiropoulos