CrackSegDiff: Diffusion Probability Model-based Multi-modal Crack Segmentation

TL;DR

CrackSegDiff introduces a novel diffusion probabilistic model that fuses grayscale and depth images for improved crack segmentation, especially in detecting shallow cracks, surpassing existing methods.

Contribution

The paper presents a new DPM-based crack segmentation approach that uniquely combines local and global features using Vm-unet, CFM, and SFCM modules, enhancing performance over prior techniques.

Findings

Outperforms state-of-the-art methods on FIND dataset

Particularly excels in shallow crack detection

Effective multi-modal data fusion improves segmentation accuracy

Abstract

Integrating grayscale and depth data in road inspection robots could enhance the accuracy, reliability, and comprehensiveness of road condition assessments, leading to improved maintenance strategies and safer infrastructure. However, these data sources are often compromised by significant background noise from the pavement. Recent advancements in Diffusion Probabilistic Models (DPM) have demonstrated remarkable success in image segmentation tasks, showcasing potent denoising capabilities, as evidenced in studies like SegDiff. Despite these advancements, current DPM-based segmentors do not fully capitalize on the potential of original image data. In this paper, we propose a novel DPM-based approach for crack segmentation, named CrackSegDiff, which uniquely fuses grayscale and range/depth images. This method enhances the reverse diffusion process by intensifying the interaction between local feature extraction via DPM and global feature extraction. Unlike traditional methods that utilize Transformers for global features, our approach employs Vm-unet to efficiently capture long-range information of the original data. The integration of features is further refined through two innovative modules: the Channel Fusion Module (CFM) and the Shallow Feature Compensation Module (SFCM). Our experimental evaluation on the three-class crack image segmentation tasks within the FIND dataset demonstrates that CrackSegDiff outperforms state-of-the-art methods, particularly excelling in the detection of shallow cracks. Code is available at https://github.com/sky-visionX/CrackSegDiff.