TY  - RPRT
TI  - MedARC: Training-Free Adaptive Redundancy Compression of Visual Tokens for 3D Medical Vision-Language Models
AU  - Yitao Zhu
AU  - Mengjun Liu
AU  - Yingji Fu
AU  - Haowen Pang
AU  - Anqi Qiu
PY  - 2026
UR  - https://arxiv.org/abs/2607.26554
ID  - 2607.26554
ER  -