LoC-Path: Learning to Compress for Pathology Multimodal Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hu, Qingqiao, Lyu, Weimin, Xu, Meilong, Qi, Kehan, Hu, Xiaoling, Gupta, Saumya, Zhou, Jiawei, Chen, Chao
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915855773728768
author Hu, Qingqiao
Lyu, Weimin
Xu, Meilong
Qi, Kehan
Hu, Xiaoling
Gupta, Saumya
Zhou, Jiawei
Chen, Chao
author_facet Hu, Qingqiao
Lyu, Weimin
Xu, Meilong
Qi, Kehan
Hu, Xiaoling
Gupta, Saumya
Zhou, Jiawei
Chen, Chao
contents Whole Slide Image (WSI) MLLMs are difficult to build and deploy because gigapixel slides induce thousands of visual tokens, while only a small fraction of regions is diagnostically relevant. Existing slide-level pathology MLLMs typically combine heavy slide-level encoders with long visual prefixes, making end-to-end slide-level development and deployment expensive under limited computational resources. We revisit this regime and show that WSI tile features are highly redundant at both global and local scales, while task-relevant evidence is sparse and query-dependent. We therefore introduce LoC-Path, a resource-efficient slide-level MLLM that compresses before fusion. LoC-Path uses a Sparse Token Merger (STM) and an MAE-pretrained resampler to replace expensive slide-level encoding with a compact latent interface, then uses a Token Importance Scorer (TIS) to select the most relevant latents and a Cross-Attention Routing Adapter (CARA) to fuse them into a few LLM decoder layers. This design lowers both multimodal tuning cost and inference-time latency/memory by avoiding heavy slide-level encoding and long visual prefixes. Extensive experiments show that LoC-Path remains competitive with prior slide-level MLLMs while making end-to-end development and deployment more practical under limited computational resources.
format Preprint
id arxiv_https___arxiv_org_abs_2512_05391
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LoC-Path: Learning to Compress for Pathology Multimodal Large Language Models
Hu, Qingqiao
Lyu, Weimin
Xu, Meilong
Qi, Kehan
Hu, Xiaoling
Gupta, Saumya
Zhou, Jiawei
Chen, Chao
Computer Vision and Pattern Recognition
Whole Slide Image (WSI) MLLMs are difficult to build and deploy because gigapixel slides induce thousands of visual tokens, while only a small fraction of regions is diagnostically relevant. Existing slide-level pathology MLLMs typically combine heavy slide-level encoders with long visual prefixes, making end-to-end slide-level development and deployment expensive under limited computational resources. We revisit this regime and show that WSI tile features are highly redundant at both global and local scales, while task-relevant evidence is sparse and query-dependent. We therefore introduce LoC-Path, a resource-efficient slide-level MLLM that compresses before fusion. LoC-Path uses a Sparse Token Merger (STM) and an MAE-pretrained resampler to replace expensive slide-level encoding with a compact latent interface, then uses a Token Importance Scorer (TIS) to select the most relevant latents and a Cross-Attention Routing Adapter (CARA) to fuse them into a few LLM decoder layers. This design lowers both multimodal tuning cost and inference-time latency/memory by avoiding heavy slide-level encoding and long visual prefixes. Extensive experiments show that LoC-Path remains competitive with prior slide-level MLLMs while making end-to-end development and deployment more practical under limited computational resources.
title LoC-Path: Learning to Compress for Pathology Multimodal Large Language Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.05391