SDDF: Specificity-Driven Dynamic Focusing for Open-Vocabulary Camouflaged Object Detection

TL;DR

This paper introduces SDDF, a method that enhances open-vocabulary camouflaged object detection by leveraging specificity-driven dynamic focusing and a new benchmark, achieving significant detection performance improvements.

Contribution

The paper proposes a novel specificity-guided regional weak alignment and dynamic focusing approach for camouflaged object detection in open-vocabulary settings, along with a new benchmark dataset.

Findings

Achieves an AP of 56.4 on the OVCOD-D benchmark.

Reduces noisy textual components with a contrastive fusion strategy.

Improves discrimination of camouflaged objects from background.

Abstract

Open-vocabulary object detection (OVOD) aims to detect known and unknown objects in the open world by leveraging text prompts. Benefiting from the emergence of large-scale vision--language pre-trained models, OVOD has demonstrated strong zero-shot generalization capabilities. However, when dealing with camouflaged objects, the detector often fails to distinguish and localize objects because the visual features of the objects and the background are highly similar. To bridge this gap, we construct a benchmark named OVCOD-D by augmenting carefully selected camouflaged object images with fine-grained textual descriptions. Due to the limited scale of available camouflaged object datasets, we adopt detectors pre-trained on large-scale object detection datasets as our baseline methods, as they possess stronger zero-shot generalization ability. In the specificity-aware sub-descriptions generated by multimodal large models, there still exist confusing and overly decorative modifiers. To mitigate such interference, we design a sub-description principal component contrastive fusion strategy that reduces noisy textual components. Furthermore, to address the challenge that the visual features of camouflaged objects are highly similar to those of their surrounding environment, we propose a specificity-guided regional weak alignment and dynamic focusing method, which aims to strengthen the detector's ability to discriminate camouflaged objects from background. Under the open-set evaluation setting, the proposed method achieves an AP of 56.4 on the OVCOD-D benchmark.