Residual-based Language Models are Free Boosters for Biomedical Imaging Tasks

Zhixin Lai, Jing Wu, Suiyao Chen, Yucheng Zhou, Naira Hovakimyan; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2024, pp. 5086-5096

In this study we uncover the unexpected efficacy of residual-based large language models (LLMs) as part of encoders for biomedical imaging tasks a domain traditionally devoid of language or textual data. The approach diverges from established methodologies by utilizing a frozen transformer block extracted from pre-trained LLMs as an innovative encoder layer for the direct processing of visual tokens. This strategy represents a significant departure from the standard multi-modal vision-language frameworks which typically hinge on language-driven prompts and inputs. We found that these LLMs could boost performance across a spectrum of biomedical imaging applications including both 2D and 3D visual classification tasks serving as plug-and-play boosters. More interestingly as a byproduct we found that the proposed framework achieved superior performance setting new state-of-the-art results on extensive standardized datasets in MedMNIST-2D and 3D. Through this work we aim to open new avenues for employing LLMs in biomedical imaging and enriching the understanding of their potential in this specialized domain.