Pretrained Equivariant Features Improve Unsupervised Landmark Discovery

Locating semantically meaningful landmark points is a crucial component of a large number of computer vision pipelines. Because of the small number of available datasets with ground truth landmark annotations, it is essential to design robust unsupervised and semi-supervised methods for landmark detection. Many of the recent unsupervised learning methods rely on the equivariance properties of landmarks to synthetic image deformations. Our work focuses on such widely used methods and sheds light on its core problem, its inability to produce equivariant intermediate convolutional features. This finding leads us to formulate a two-step unsupervised approach that overcomes this challenge by first learning powerful pixel-based features that are then exploited in a second step within standard landmark detection approaches. Our methodology produces state-of-the-art results on several benchmarks such as the BBC Pose dataset, the Cat-head dataset and performs comparably in other situations.