Improving Long-Tailed Object Detection with Balanced Group Softmax and Metric Learning

TL;DR

This paper advances long-tailed object detection by enhancing the Balanced Group Softmax framework and integrating metric learning with k-NN inference, achieving state-of-the-art results on the LVISv1 dataset.

Contribution

It introduces improvements to the BAGS framework and applies metric learning with k-NN for better rare class detection in long-tailed datasets.

Findings

Achieved a new state-of-the-art mAP of 24.5% on LVISv1.

Enhanced class imbalance mitigation with improved BAGS.

Demonstrated the effectiveness of metric learning and k-NN for rare class detection.

Abstract

Object detection has been widely explored for class-balanced datasets such as COCO. However, real-world scenarios introduce the challenge of long-tailed distributions, where numerous categories contain only a few instances. This inherent class imbalance biases detection models towards the more frequent classes, degrading performance on rare categories. In this paper, we tackle the problem of long-tailed 2D object detection using the LVISv1 dataset, which consists of 1,203 categories and 164,000 images. We employ a two-stage Faster R-CNN architecture and propose enhancements to the Balanced Group Softmax (BAGS) framework to mitigate class imbalance. Our approach achieves a new state-of-the-art performance with a mean Average Precision (mAP) of 24.5%, surpassing the previous benchmark of 24.0%. Additionally, we hypothesize that tail class features may form smaller, denser clusters within the feature space of head classes, making classification challenging for regression-based classifiers. To address this issue, we explore metric learning to produce feature embeddings that are both well-separated across classes and tightly clustered within each class. For inference, we utilize a k-Nearest Neighbors (k-NN) approach to improve classification performance, particularly for rare classes. Our results demonstrate the effectiveness of these methods in advancing long-tailed object detection.