PACF: Prototype Augmented Compact Features for Improving Domain Adaptive Object Detection

2025-01-15

AI Summary

To address class-conditional feature distribution divergence and class-center shift caused by domain shift in cross-domain object detection, this paper proposes a prototype-enhanced compact feature framework. Methodologically: (1) we derive a prototype cross-entropy loss from the evidence lower bound (ELBO) theory, explicitly enforcing intra-class compactness on the target domain; (2) we design a mutual regularization mechanism between a linear classifier and a prototype-based classifier to jointly suppress intra-class variance and inter-class mean shift; (3) we introduce RoI-level feature distribution calibration. Evaluated on multiple domain-adaptive detection benchmarks, our method achieves state-of-the-art performance, significantly reducing target-domain class-conditional variance (average ↓18.7%) and cross-domain class-center shift (average ↓23.4%), thereby validating the efficacy of theory-driven loss design and dual-classifier collaborative optimization.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Economics, Online Markets and Human Computation: Humans versus LLMs for data annotation and labeling
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

In recent years, there has been significant advancement in object detection. However, applying off-the-shelf detectors to a new domain leads to significant performance drop, caused by the domain gap. These detectors exhibit higher-variance class-conditional distributions in the target domain than that in the source domain, along with mean shift. To address this problem, we propose the Prototype Augmented Compact Features (PACF) framework to regularize the distribution of intra-class features. Specifically, we provide an in-depth theoretical analysis on the lower bound of the target features-related likelihood and derive the prototype cross entropy loss to further calibrate the distribution of target RoI features. Furthermore, a mutual regularization strategy is designed to enable the linear and prototype-based classifiers to learn from each other, promoting feature compactness while enhancing discriminability. Thanks to this PACF framework, we have obtained a more compact cross-domain feature space, within which the variance of the target features' class-conditional distributions has significantly decreased, and the class-mean shift between the two domains has also been further reduced. The results on different adaptation settings are state-of-the-art, which demonstrate the board applicability and effectiveness of the proposed approach.

Problem

• Object Recognition
• Environmental Variability
• Accuracy Degradation

Innovation

• PACF
• Domain Adaptation
• Cross-Environment Object Recognition