Logsage: Log-Based Saliency for Guided Encoding in Robust Nuclei Segmentation of Immunofluorescence Histology Images

The tumor microenvironment (TME) is critical in cancer progression, development, and treatment response. However, its complex cellular architecture (e.g., cell type, organization) presents significant challenges for accurate immunofluorescence (IF) image segmentation. We introduce <b>LoGSAGE</b>-Net (<b>LoG</b>-based <b>SA</b>liency for <b>G</b>uided <b>E</b>ncoding), which couples a Swin Transformer with the encoded response from Laplacian of Gaussian (LoG) on multiple scales. The loss function incorporates two deformation metrics, combining the Dice- and curvature alignment loss. The model is applied to a large cohort of preclinical data and has shown an improved performance over the state-of-the-art methods. The proposed model achieved a Dice score of 94.92% and a Panoptic Quality (PQ) score of 81%. This model supports robust profiling of the TME for sensitive assays.