Plug-and-Play DISep: Separating Dense Instances for Scene-to-Pixel Weakly-Supervised Change Detection in High-Resolution Remote Sensing Images

2025-01-09

AI Summary

Existing weakly supervised change detection (WSCD) methods struggle to accurately localize and separate densely packed change instances in high-resolution remote sensing imagery, often misidentifying unchanged regions as changed—leading to inaccurate estimation of both the number and spatial locations of changes. To address this, we propose DISep, a plug-and-play Dense Instance Separation framework featuring a novel iterative “localize–retrieve–separate” mechanism. DISep achieves feature disentanglement of change instances using only scene-level labels—without requiring pixel-level annotations. It integrates high-pass-filtered class activation maps, connected-component-guided retrieval, instance-aware separation loss, and embedding-space consistency constraints, and is compatible with both CNN- and Transformer-based backbones. Evaluated on five standard benchmarks—including LEVIR-CD—DISep consistently improves the performance of seven state-of-the-art WSCD methods, achieving new SOTA results. It incurs zero inference overhead and further enhances fully supervised methods through transferability.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Semi-Supervised Learning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments

Abstract

Existing Weakly-Supervised Change Detection (WSCD) methods often encounter the problem of"instance lumping"under scene-level supervision, particularly in scenarios with a dense distribution of changed instances (i.e., changed objects). In these scenarios, unchanged pixels between changed instances are also mistakenly identified as changed, causing multiple changes to be mistakenly viewed as one. In practical applications, this issue prevents the accurate quantification of the number of changes. To address this issue, we propose a Dense Instance Separation (DISep) method as a plug-and-play solution, refining pixel features from a unified instance perspective under scene-level supervision. Specifically, our DISep comprises a three-step iterative training process: 1) Instance Localization: We locate instance candidate regions for changed pixels using high-pass class activation maps. 2) Instance Retrieval: We identify and group these changed pixels into different instance IDs through connectivity searching. Then, based on the assigned instance IDs, we extract corresponding pixel-level features on a per-instance basis. 3) Instance Separation: We introduce a separation loss to enforce intra-instance pixel consistency in the embedding space, thereby ensuring separable instance feature representations. The proposed DISep adds only minimal training cost and no inference cost. It can be seamlessly integrated to enhance existing WSCD methods. We achieve state-of-the-art performance by enhancing {three Transformer-based and four ConvNet-based methods} on the LEVIR-CD, WHU-CD, DSIFN-CD, SYSU-CD, and CDD datasets. Additionally, our DISep can be used to improve fully-supervised change detection methods. Code is available at https://github.com/zhenghuizhao/Plug-and-Play-DISep-for-Change-Detection.

Problem

• Change Detection
• High Resolution Imagery
• Object Recognition

Innovation

• DISep
• Change Detection
• High-resolution Imagery