Learning to Navigate for Fine-grained Classification

TL;DR

This paper introduces NTS-Net, a multi-agent self-supervised framework that localizes informative regions for fine-grained classification without bounding-box annotations, achieving state-of-the-art results.

Contribution

It proposes a novel multi-agent training paradigm with a Navigator, Teacher, and Scrutinizer to improve fine-grained classification by localizing discriminative regions.

Findings

Achieves state-of-the-art accuracy on benchmark datasets.

Effectively localizes informative regions without bounding-box annotations.

End-to-end trainable model with high interpretability.

Abstract

Fine-grained classification is challenging due to the difficulty of finding discriminative features. Finding those subtle traits that fully characterize the object is not straightforward. To handle this circumstance, we propose a novel self-supervision mechanism to effectively localize informative regions without the need of bounding-box/part annotations. Our model, termed NTS-Net for Navigator-Teacher-Scrutinizer Network, consists of a Navigator agent, a Teacher agent and a Scrutinizer agent. In consideration of intrinsic consistency between informativeness of the regions and their probability being ground-truth class, we design a novel training paradigm, which enables Navigator to detect most informative regions under the guidance from Teacher. After that, the Scrutinizer scrutinizes the proposed regions from Navigator and makes predictions. Our model can be viewed as a multi-agent cooperation, wherein agents benefit from each other, and make progress together. NTS-Net can be trained end-to-end, while provides accurate fine-grained classification predictions as well as highly informative regions during inference. We achieve state-of-the-art performance in extensive benchmark datasets.