cs.CVNov 5, 2026

VoluCompose: Constructing Compact Volumetric Visual Representations for Medical Image Understanding

Authors: Zanting Ye, Yunrui Lin, Shengyuan Liu, Jing Hao, Chen Hu, Yefeng Zheng, Lijun Lu

Organizations: School of Biomedical Engineering and Guangdong Provincial Key Laboratory of Medical Image Processing, Southern Medical University, Guangzhou 510515, China · Medical Artificial Intelligence Laboratory, Westlake University, Hangzhou 310030, China · Department of Electronic Engineering, The Chinese University of Hong Kong, Hong Kong SAR, China · Faculty of Dentistry, The University of Hong Kong, Hong Kong SAR, China

Abstract

Medical vision--language models can interpret volumetric images, but processing full volumes is costly because hundreds of slices produce long, redundant visual-token sequences. Existing compression methods often couple content selection with budget determination. We observe that neighboring slices share visual features and that attention concentration varies across visual encoders. We propose VoluCompose, a training-free, model-agnostic framework that separates what visual content to retain from how many tokens to allocate. Its hyperbolic token composition module groups contiguous slices into ordered support units and uses the Lorentz distance to account for feature magnitude and direction during token allocation, anchor selection, and refinement. Its capacity control module uses visual-encoder attention entropy to allocate more tokens when attention is more dispersed. We evaluate VoluCompose on two tasks across three datasets and four backbones. Across nine calibrated settings, VoluCompose attains the highest mean quality--retention hypervolume among the compared methods. It achieves a mean relative quality of 121.68% against the uncompressed references while retaining 22.78% of visual tokens on average. Compression remains robust even at a 10% token budget. Across settings, the lowest ROUGE-L ratio to the uncompressed reference is 96.4% for VoluCompose. The corresponding minima for the compression baselines range from 15.0% to 70.5%. Code is available at https://github.com/yezanting/VoluCompose.

Figures & tables

Explore similar work

CardsList
  1. CARVE: Cross-Slice Anisotropic Reallocation of Visual Evidence for Efficient 3D Medical Volume Understanding

    Aug 5, 2026Zhenyu Yi, Qiang Hu, Zhenhao Li +33D Medical ImagingMedical Visual Question Answering

  2. MedARC: Training-Free Adaptive Redundancy Compression of Visual Tokens for 3D Medical Vision-Language Models

    Jul 29, 2026Yitao Zhu, Mengjun Liu, Yingji Fu +2Efficient VLM InferenceMedical VLMs

  3. Foveated Compression: Selective High-Resolution Preservation for Token-Efficient VLMs

    Oct 6, 2026Donghyun Han, Jangho Park, Yuseok BaeEfficient VLM InferenceVisual Token Pruning