Resolving Class Imbalance in Object Detection with Weighted Cross Entropy Losses

TL;DR

This paper addresses class imbalance in object detection by applying weighted cross entropy variants, improving detection accuracy for minority classes in highly imbalanced datasets like BDD100K.

Contribution

It introduces and evaluates weighted cross entropy loss functions, such as Focal Loss and Class-Balanced Loss, to enhance class-wise detection performance in imbalanced datasets.

Findings

Weighted loss functions improve minority class detection

Experiments on BDD100K show better class-wise accuracy

Weighted losses outperform standard cross entropy in imbalanced scenarios

Abstract

Object detection is an important task in computer vision which serves a lot of real-world applications such as autonomous driving, surveillance and robotics. Along with the rapid thrive of large-scale data, numerous state-of-the-art generalized object detectors (e.g. Faster R-CNN, YOLO, SSD) were developed in the past decade. Despite continual efforts in model modification and improvement in training strategies to boost detection accuracy, there are still limitations in performance of detectors when it comes to specialized datasets with uneven object class distributions. This originates from the common usage of Cross Entropy loss function for object classification sub-task that simply ignores the frequency of appearance of object class during training, and thus results in lower accuracies for object classes with fewer number of samples. Class-imbalance in general machine learning has been widely studied, however, little attention has been paid on the subject of object detection. In this paper, we propose to explore and overcome such problem by application of several weighted variants of Cross Entropy loss, for examples Balanced Cross Entropy, Focal Loss and Class-Balanced Loss Based on Effective Number of Samples to our object detector. Experiments with BDD100K (a highly class-imbalanced driving database acquired from on-vehicle cameras capturing mostly Car-class objects and other minority object classes such as Bus, Person and Motor) have proven better class-wise performances of detector trained with the afore-mentioned loss functions.