SGNet: A Super-class Guided Network for Image Classification and Object Detection

TL;DR

SGNet leverages high-level super-class information to guide image classification and object detection, improving accuracy by integrating semantic hierarchies into the prediction process.

Contribution

The paper introduces a super-class guided network (SGNet) that uses hierarchical class annotations and a super-class branch to enhance prediction accuracy.

Findings

SGNet outperforms baseline models on CIFAR-100 and MS COCO datasets.

Two inference strategies, TSI and DI, effectively utilize super-class information.

Experimental results show significant performance improvements in classification and detection.

Abstract

Most classification models treat different object classes in parallel and the misclassifications between any two classes are treated equally. In contrast, human beings can exploit high-level information in making a prediction of an unknown object. Inspired by this observation, the paper proposes a super-class guided network (SGNet) to integrate the high-level semantic information into the network so as to increase its performance in inference. SGNet takes two-level class annotations that contain both super-class and finer class labels. The super-classes are higher-level semantic categories that consist of a certain amount of finer classes. A super-class branch (SCB), trained on super-class labels, is introduced to guide finer class prediction. At the inference time, we adopt two different strategies: Two-step inference (TSI) and direct inference (DI). TSI first predicts the super-class and then makes predictions of the corresponding finer class. On the other hand, DI directly generates predictions from the finer class branch (FCB). Extensive experiments have been performed on CIFAR-100 and MS COCO datasets. The experimental results validate the proposed approach and demonstrate its superior performance on image classification and object detection.