Efficient Medical Vision-Language Alignment Through Adapting Masked Vision Models

计算机视觉 人工智能 计算机科学 机器视觉 图像配准 医学影像学 图像(数学)
作者
Chenyu Lian,Hong-Yu Zhou,Dongyun Liang,Jing Qin,Liansheng Wang
出处
期刊:IEEE Transactions on Medical Imaging [Institute of Electrical and Electronics Engineers]
卷期号:44 (11): 4499-4510 被引量:1
标识
DOI:10.1109/tmi.2025.3575853
摘要

Medical vision-language alignment through cross-modal contrastive learning shows promising performance in image-text matching tasks, such as retrieval and zero-shot classification. However, conventional cross-modal contrastive learning (CLIP-based) methods suffer from suboptimal visual representation capabilities, which also limits their effectiveness in vision-language alignment. In contrast, although the models pretrained via multimodal masked modeling struggle with direct cross-modal matching, they excel in visual representation. To address this contradiction, we propose ALTA (ALign Through Adapting), an efficient medical vision-language alignment method that utilizes only about 8% of the trainable parameters and less than 1/5 of the computational consumption required for masked record modeling. ALTA achieves superior performance in vision-language matching tasks like retrieval and zero-shot classification by adapting the pretrained vision model from masked record modeling. Additionally, we integrate temporal-multiview radiograph inputs to enhance the information consistency between radiographs and their corresponding descriptions in reports, further improving the vision-language alignment. Experimental evaluations show that ALTA outperforms the best-performing counterpart by over 4% absolute points in text-to-image accuracy and approximately 6% absolute points in image-to-text retrieval accuracy. The adaptation of vision-language models during efficient alignment also promotes better vision and language understanding. Code is publicly available at https://github.com/DopamineLcy/ALTA.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
科研通AI6.4应助Autumn采纳,获得10
刚刚
雪儿完成签到,获得积分10
刚刚
英俊的铭应助千纸鹤采纳,获得10
1秒前
TANGchenran发布了新的文献求助20
2秒前
3秒前
5秒前
5秒前
珊任宗珊完成签到,获得积分10
6秒前
6秒前
山野的雾发布了新的文献求助10
6秒前
Elia完成签到,获得积分10
7秒前
王WW完成签到,获得积分10
7秒前
9秒前
科研通AI6.2应助loooi采纳,获得10
9秒前
SciGPT应助2064777494采纳,获得10
11秒前
yulk发布了新的文献求助10
11秒前
12秒前
12秒前
liufan完成签到 ,获得积分10
12秒前
尹依依发布了新的文献求助10
12秒前
小小小乐完成签到 ,获得积分10
12秒前
13秒前
阿吉利拉完成签到,获得积分20
13秒前
逃离乌托邦完成签到,获得积分10
13秒前
yy发布了新的文献求助10
16秒前
16秒前
炙热的元彤完成签到,获得积分10
17秒前
17秒前
ccc发布了新的文献求助10
17秒前
丘比特应助veggieg采纳,获得10
17秒前
wz完成签到,获得积分10
17秒前
顾矜应助veggieg采纳,获得10
17秒前
科研通AI6.4应助veggieg采纳,获得10
18秒前
万能图书馆应助veggieg采纳,获得10
18秒前
研友_VZG7GZ应助veggieg采纳,获得10
18秒前
贪玩的秋柔应助veggieg采纳,获得10
18秒前
18秒前
今后应助veggieg采纳,获得10
18秒前
桐桐应助veggieg采纳,获得10
18秒前
芮rich完成签到,获得积分10
19秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
Essentials of Carbohydrate Chemistry and Biochemistry, 4th Edition 800
Organizational Behavior 510
Management and the Arts 510
Matrix Methods in Data Mining and Pattern Recognition Second Edition 510
CLSI VET01S-2024 Performance Standards for Antimicrobial Disk and Dilution Susceptibility Tests for Bacteria Isolated From Animals (7th Ed) 500
A Case Study on Hotels as Noncongregate Emergency Living Accommodations for Returning Citizens 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 纳米技术 计算机科学 化学工程 工程类 有机化学 物理 复合材料 生物化学 内科学 细胞生物学 基因 遗传学 免疫学 冶金 光电子学 癌症研究
热门帖子
关注 科研通微信公众号,转发送积分 7765161
求助须知:如何正确求助?哪些是违规求助? 9309488
关于积分的说明 20311129
捐赠科研通 7349938
什么是DOI,文献DOI怎么找? 3314730
关于科研通互助平台的介绍 2464132
邀请新用户注册赠送积分活动 2329189