Learning Dense Object Descriptors from Multiple Views for Low-shot Category Generalization
This addresses the challenge of low-shot category recognition in computer vision, offering a self-supervised method that reduces reliance on labeled data, though it is incremental as it builds on existing multi-view and self-supervised techniques.
The paper tackles the problem of learning dense object descriptors for low-shot category generalization without category labels, by proposing Deep Object Patch Encodings (DOPE) trained from multiple views using self-supervision, and finds it outperforms supervised and self-supervised baselines in low-shot classification.
A hallmark of the deep learning era for computer vision is the successful use of large-scale labeled datasets to train feature representations for tasks ranging from object recognition and semantic segmentation to optical flow estimation and novel view synthesis of 3D scenes. In this work, we aim to learn dense discriminative object representations for low-shot category recognition without requiring any category labels. To this end, we propose Deep Object Patch Encodings (DOPE), which can be trained from multiple views of object instances without any category or semantic object part labels. To train DOPE, we assume access to sparse depths, foreground masks and known cameras, to obtain pixel-level correspondences between views of an object, and use this to formulate a self-supervised learning task to learn discriminative object patches. We find that DOPE can directly be used for low-shot classification of novel categories using local-part matching, and is competitive with and outperforms supervised and self-supervised learning baselines. Code and data available at https://github.com/rehg-lab/dope_selfsup.