Fidelity-Imposed Displacement Editing for the Learn2Reg 2024 SHG-BF Challenge

Wang, Jiacheng, Chen, Xiang, Hu, Renjiu, Wang, Rongguang, Liu, Min, Wang, Yaonan, Wang, Jiazheng, Li, Hao, Zhang, Hang

arXiv.org Artificial Intelligence 

To address these challenges, we propose a novel SHG-BF Co-examination of second-harmonic generation (SHG) and multimodal registration method with the following key contributions: bright-field (BF) microscopy enables the differentiation of tissue components and collagen fibers, aiding the analysis of human 1. Batch-wise contrastive loss (B-NCE): We introduce a breast and pancreatic cancer tissues. However, large discrepancies batch-wise noise contrastive estimation loss to effectively between SHG and BF images pose challenges for capture shared features between SHG and BF images.