CoDS: Enhancing Collaborative Perception in Heterogeneous Scenarios via Domain Separation

2025-10-15

AI Summary

In autonomous driving heterogeneous collaborative perception, feature inconsistency arises from encoder heterogeneity across agents, and existing methods—relying on feature alignment and Transformer-based domain adaptation—are vulnerable to domain noise and suffer from low inference efficiency. Method: This paper proposes a domain-separation-driven lightweight collaborative perception framework. It introduces an encoder-specificity/generalization disentanglement module to separate domain-dependent and task-relevant information; employs a lightweight convolutional Local Spatial-Channel Refinement (LSCR) for spatial-channel alignment; adopts the Distribution Alignment via Domain Separation (DADS) module for distribution alignment; and applies the Domain-Agnostic Mutual Information (DAMI) loss to maximize mutual information. A fully convolutional architecture ensures efficient mobile deployment. Contribution/Results: Experiments demonstrate that the method significantly mitigates heterogeneous feature bias, achieving superior trade-offs between detection accuracy and inference speed.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Application Domains: Mobility, Driving & Flight

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Collaborative perception has been proven to improve individual perception in autonomous driving through multi-agent interaction. Nevertheless, most methods often assume identical encoders for all agents, which does not hold true when these models are deployed in real-world applications. To realize collaborative perception in actual heterogeneous scenarios, existing methods usually align neighbor features to those of the ego vehicle, which is vulnerable to noise from domain gaps and thus fails to address feature discrepancies effectively. Moreover, they adopt transformer-based modules for domain adaptation, which causes the model inference inefficiency on mobile devices. To tackle these issues, we propose CoDS, a Collaborative perception method that leverages Domain Separation to address feature discrepancies in heterogeneous scenarios. The CoDS employs two feature alignment modules, i.e., Lightweight Spatial-Channel Resizer (LSCR) and Distribution Alignment via Domain Separation (DADS). Besides, it utilizes the Domain Alignment Mutual Information (DAMI) loss to ensure effective feature alignment. Specifically, the LSCR aligns the neighbor feature across spatial and channel dimensions using a lightweight convolutional layer. Subsequently, the DADS mitigates feature distribution discrepancy with encoder-specific and encoder-agnostic domain separation modules. The former removes domain-dependent information and the latter captures task-related information. During training, the DAMI loss maximizes the mutual information between aligned heterogeneous features to enhance the domain separation process. The CoDS employs a fully convolutional architecture, which ensures high inference efficiency. Extensive experiments demonstrate that the CoDS effectively mitigates feature discrepancies in heterogeneous scenarios and achieves a trade-off between detection accuracy and inference efficiency.

Problem

• Addressing feature discrepancies in heterogeneous collaborative perception scenarios
• Improving model inference efficiency for mobile device deployment
• Mitigating noise from domain gaps in multi-agent perception systems

Innovation

• Lightweight Spatial-Channel Resizer for feature alignment
• Domain Separation modules for distribution discrepancy mitigation
• Fully convolutional architecture ensuring high inference efficiency