Learning Pyramid Representations from Gigapixel Histopathological Images

Weiyi Wu, Xingjian Diao, Chunhui Zhang, Chongyang Gao, Xinwen Xu, Siting Li, Jiang Gui

Keywords: Computer Vision, Transformer

TL;DR: Learning Pyramid Representations from Gigapixel Histopathological Images

Abstract: Whole slide images (WSIs) pose fundamental computational challenges due to their gigapixel resolution and the sparse distribution of informative regions. Existing approaches often treat image patches independently—discarding spatial structure—or reshape them in ways that distort spatial context, thereby obscuring the hierarchical pyramid representations intrinsic to WSIs. We introduce Sparse Pyramid Attention Networks (SPAN), a hierarchical framework that preserves spatial relationships while efficiently allocating computation to informative regions. SPAN constructs multi-scale representations directly from single-scale inputs, enabling precise WSI modeling without sacrificing efficiency. We demonstrate SPAN’s versatility through two variants: SPAN-MIL for slide classification and SPAN-UNet for segmentation. Comprehensive evaluations across multiple public datasets show that SPAN captures the hierarchical structure and contextual relationships that existing methods fail to model. Our results provide clear evidence that architectural inductive biases and hierarchical representations enhance both slide-level and patch-level performance. By overcoming long-standing computational barriers, SPAN establishes a new paradigm for computational pathology and reveals foundational design principles for large-scale medical image analysis.

Primary Area: applications to computer vision, audio, language, and other modalities

Submission Number: 24460