SOURCE-LINKED INTELLIGENCE

B-MIM: Biased Masked Image Modeling for Generalizable Segmentation of Fine-Grained Anatomical Structures

arXiv · AI, language, vision and robotics · article · Aug 25, 2026 · UTC

Self-supervised pretraining enables transferable representations for medical imaging, yet most CT encoders remain biased toward coarse semantic understanding, limiting their sensitivity to fine-grained anatomical structures such as vessels or small tumors. In this paper, we introduce Biased Masked Image Modeling (B-MIM), a modification of the iBOT objective that stochastically reduces global semantic alignment to prioritize local patch reconstruction. This bias encourages the encoder to capture high-frequency morphological details and structural continuity. We curate a multi-institutional CT a

Read original source ↗ Open in workspace

recordType

paper

region

Global