Open access
Jul 2026
An Improved Generative Adversarial Network for Footprint Image Segmentation
Findings indicate that, under the controlled acquisition conditions of the aFoot_2025 dataset, CPCA-based feature calibration and dual-scale discrimination may improve segmentation-mask quality and provide a possible benefit for subsequent anthropometric analysis.
Dongliang Yang, Chang-Jiang Song, Xianglei Xing
· Electronics · 0 citations