Utilizing Grounded SAM for self-supervised frugal camouflaged human detection

TL;DR

This paper explores self-supervised and frugal learning methods for camouflaged human detection, fine-tuning existing models with limited labeled data, and demonstrates comparable performance to fully supervised approaches.

Contribution

It introduces self-supervised pseudo-labeling techniques for camouflaged human detection, reducing reliance on large labeled datasets and adapting existing models effectively.

Findings

Self-supervised methods achieve similar performance to supervised ones.

Frugal transfer learning effectively fine-tunes models for camouflaged human detection.

Pseudo-labeling approaches are viable for reducing annotation effort.

Abstract

Visually detecting camouflaged objects is a hard problem for both humans and computer vision algorithms. Strong similarities between object and background appearance make the task significantly more challenging than traditional object detection or segmentation tasks. Current state-of-the-art models use either convolutional neural networks or vision transformers as feature extractors. They are trained in a fully supervised manner and thus need a large amount of labeled training data. In this paper, both self-supervised and frugal learning methods are introduced to the task of Camouflaged Object Detection (COD). The overall goal is to fine-tune two COD reference methods, namely SINet-V2 and HitNet, pre-trained for camouflaged animal detection to the task of camouflaged human detection. Therefore, we use the public dataset CPD1K that contains camouflaged humans in a forest environment. We create a strong baseline using supervised frugal transfer learning for the fine-tuning task. Then, we analyze three pseudo-labeling approaches to perform the fine-tuning task in a self-supervised manner. Our experiments show that we achieve similar performance by pure self-supervision compared to fully supervised frugal learning.