Background Fades, Foreground Leads: Curriculum-Guided Background Pruning for Efficient Foreground-Centric Collaborative Perception

2025-10-22

AI Summary

To address context information loss in cooperative perception under vehicle-to-vehicle bandwidth constraints—caused by discarding background features—this paper proposes a foreground-dominated feature sharing framework. Methodologically, it introduces (1) a curriculum-based background pruning mechanism that progressively masks background regions during training, guiding the model to compress and internalize essential environmental context into foreground features; and (2) foreground-centered feature extraction coupled with context distillation to enhance the semantic completeness of foreground representations. By eliminating background transmission, the method significantly improves feature compression efficiency and reconstruction fidelity. Evaluated on both synthetic and real-world benchmarks under diverse bandwidth constraints, it consistently outperforms existing state-of-the-art methods, achieving average detection accuracy gains of 3.2%–5.8%. These results validate the effectiveness and generalizability of context-enhanced foreground sharing for bandwidth-limited cooperative perception.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Feature Construction/Reformulation
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Location- and context-aware Web and WoT applications and services
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Collaborative perception enhances the reliability and spatial coverage of autonomous vehicles by sharing complementary information across vehicles, offering a promising solution to long-tail scenarios that challenge single-vehicle perception. However, the bandwidth constraints of vehicular networks make transmitting the entire feature map impractical. Recent methods, therefore, adopt a foreground-centric paradigm, transmitting only predicted foreground-region features while discarding the background, which encodes essential context. We propose FadeLead, a foreground-centric framework that overcomes this limitation by learning to encapsulate background context into compact foreground features during training. At the core of our design is a curricular learning strategy that leverages background cues early on but progressively prunes them away, forcing the model to internalize context into foreground representations without transmitting background itself. Extensive experiments on both simulated and real-world benchmarks show that FadeLead outperforms prior methods under different bandwidth settings, underscoring the effectiveness of context-enriched foreground sharing.

Problem

• Transmitting entire feature maps is impractical due to bandwidth constraints
• Foreground-centric methods discard essential background context information
• Learning to encapsulate background context into compact foreground features

Innovation

• Encapsulates background context into foreground features
• Uses curricular learning to progressively prune background
• Transmits only context-enriched foreground representations