Contrastive Learning for OOD in Object detection

TL;DR

This paper explores using contrastive learning with human bias to improve class similarity ranking and representation efficiency in object detection, achieving comparable results to supervised methods and discussing limitations in out-of-distribution detection.

Contribution

It introduces a ranking-based contrastive learning approach incorporating human bias for better class representation in object detection.

Findings

Comparable performance to supervised contrastive learning in image classification and object detection

Incorporating human bias can enhance representation learning

Discusses limitations in out-of-distribution detection

Abstract

Contrastive learning is commonly applied to self-supervised learning, and has been shown to outperform traditional approaches such as the triplet loss and N-pair loss. However, the requirement of large batch sizes and memory banks has made it difficult and slow to train. Recently, Supervised Contrasative approaches have been developed to overcome these problems. They focus more on learning a good representation for each class individually, or between a cluster of classes. In this work we attempt to rank classes based on similarity using a user-defined ranking, to learn an efficient representation between all classes. We observe how incorporating human bias into the learning process could improve learning representations in the parameter space. We show that our results are comparable to Supervised Contrastive Learning for image classification and object detection, and discuss it's shortcomings in OOD Detection