Style Evolving along Chain-of-Thought for Unknown-Domain Object Detection

2025-03-13

AI Summary

This work addresses the weak generalization of single-domain generalized object detection (Single-DGOD) models under previously unseen, complex multi-style domains (e.g., rainy night). To tackle this, we propose a novel Chain-of-Thought (CoT)-based progressive style modeling paradigm. Unlike conventional one-step textual prompting, our method introduces a dynamic evolution mechanism for style descriptions along the reasoning chain, achieved through multi-step vision-language prompting, iterative refinement of style semantics, and cross-domain feature distribution expansion. This enables stepwise disentanglement and synergistic enhancement of composite styles. Evaluated on five challenging adverse-weather scenarios and the Real-to-Art benchmark, our approach significantly outperforms existing state-of-the-art methods, achieving substantial gains in both detection accuracy and robustness on unknown domains.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Natural Language Processing: Sentiment Analysis, Stylistic Analysis, and Argument Mining

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Large pretrained models with web data
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Recently, a task of Single-Domain Generalized Object Detection (Single-DGOD) is proposed, aiming to generalize a detector to multiple unknown domains never seen before during training. Due to the unavailability of target-domain data, some methods leverage the multimodal capabilities of vision-language models, using textual prompts to estimate cross-domain information, enhancing the model's generalization capability. These methods typically use a single textual prompt, often referred to as the one-step prompt method. However, when dealing with complex styles such as the combination of rain and night, we observe that the performance of the one-step prompt method tends to be relatively weak. The reason may be that many scenes incorporate not just a single style but a combination of multiple styles. The one-step prompt method may not effectively synthesize combined information involving various styles. To address this limitation, we propose a new method, i.e., Style Evolving along Chain-of-Thought, which aims to progressively integrate and expand style information along the chain of thought, enabling the continual evolution of styles. Specifically, by progressively refining style descriptions and guiding the diverse evolution of styles, this approach enables more accurate simulation of various style characteristics and helps the model gradually learn and adapt to subtle differences between styles. Additionally, it exposes the model to a broader range of style features with different data distributions, thereby enhancing its generalization capability in unseen domains. The significant performance gains over five adverse-weather scenarios and the Real to Art benchmark demonstrate the superiorities of our method.

Problem

• Generalizing object detection to unseen domains using multimodal vision-language models.
• Improving performance in complex, multi-style scenarios like rain and night.
• Enhancing model generalization through progressive style integration and evolution.

Innovation

• Progressive style integration via Chain-of-Thought
• Refining style descriptions for diverse evolution
• Enhancing generalization in unseen domains