Pose Aware Fine Grained Visual Classification

[PDF]

Abstract

We focus on the problem of fine-grained visual classification (FGVC). We posit that unreasonable effectiveness of the state-of-the-art in this area is because of similar object cat- egories present in the ImageNet dataset, which allows such models to be pretrained on a much larger set of samples and learn generic features for those object categories. We observe an important and often ignored additional structure present in an FGVC problem: the objects are captured from a small set of viewing angles only. We notice that subtle differences be- tween object categories are difficult to pick from an arbitrary angle but easier to identify from a similar pose. We show in this paper that training specialized pose experts, focusing on classification from a single, fixed pose, and combining them in an ensemble style framework successfully exploits the structure in the problem. We demonstrate the effective- ness of the proposed approach on the benchmark Stanford Cars, FGVC-Aircrafts, and DeepFashion datasets. To high- light the contribution when the target category features may not be available in a pretrained network, we test on footwear class. We contribute a new 1000 object, 12 category footwear dataset, each object captured from 4 different poses and show significant improvement on this dataset.