Information-Bottleneck Driven Binary Neural Network for Change Detection

2025-07-04

AI Summary

Conventional binarization methods directly quantize weights and activations, severely degrading the representational capacity and feature discriminability of change detection models, resulting in substantially lower accuracy compared to full-precision counterparts. Method: We propose BiCD, the first binary neural network specifically designed for change detection. Grounded in information bottleneck theory, BiCD introduces a compact, learnable module to approximate mutual information and jointly optimizes reconstruction loss and change detection loss. It employs an end-to-end dual-task co-training framework to simultaneously preserve structural fidelity and enhance discriminative capability. Contribution/Results: Evaluated on diverse street-scene and remote-sensing datasets, BiCD significantly outperforms existing binarized approaches, markedly narrowing the performance gap with full-precision models. It establishes, for the first time, a new state-of-the-art benchmark for binary neural networks in change detection.

Technology Category

• Computer Vision: Representation Learning for Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Models for Web evolution

Abstract

In this paper, we propose Binarized Change Detection (BiCD), the first binary neural network (BNN) designed specifically for change detection. Conventional network binarization approaches, which directly quantize both weights and activations in change detection models, severely limit the network's ability to represent input data and distinguish between changed and unchanged regions. This results in significantly lower detection accuracy compared to real-valued networks. To overcome these challenges, BiCD enhances both the representational power and feature separability of BNNs, improving detection performance. Specifically, we introduce an auxiliary objective based on the Information Bottleneck (IB) principle, guiding the encoder to retain essential input information while promoting better feature discrimination. Since directly computing mutual information under the IB principle is intractable, we design a compact, learnable auxiliary module as an approximation target, leading to a simple yet effective optimization strategy that minimizes both reconstruction loss and standard change detection loss. Extensive experiments on street-view and remote sensing datasets demonstrate that BiCD establishes a new benchmark for BNN-based change detection, achieving state-of-the-art performance in this domain.

Problem

• Improving binary neural networks for change detection accuracy
• Enhancing feature separability in binarized change detection models
• Overcoming limitations of conventional network binarization approaches

Innovation

• Binary neural network for change detection
• Information Bottleneck principle enhances feature discrimination
• Learnable auxiliary module approximates mutual information