Integrating Crack Causal Augmentation Framework and Dynamic Binary Threshold for imbalanced crack instance segmentation

Open Webpage

Qin Lei, Jiang Zhong, Chen Wang, Xue Li

Published: 01 Jan 2024, Last Modified: 13 Nov 2024Expert Syst. Appl. 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Crack Causal Augmentation Framework (CCAF) alleviates data imbalance in crack instance segmentation.•Dynamic Binary Threshold (DBT) branch improves crack instance segmentation accuracy.•CCAF and DBT synergistically enhance crack instance segmentation performance.•Comprehensive evaluation metrics provided for crack instance segmentation.•Comprehensive evaluation on 4 crack datasets confirms effectiveness of CCAF and DBT for segmentation.