已入深夜,您辛苦了!由于当前在线用户较少,发布求助请尽量完整的填写文献信息,科研通机器人24小时在线,伴您度过漫漫科研夜!祝你早点完成任务,早点休息,好梦!

Boosting ViT-based MRI Reconstruction from the Perspectives of Frequency Modulation, Spatial Purification, and Scale Diversification

多元化(营销策略) Boosting(机器学习) 比例(比率) 空间调制 调制(音乐) 计算机科学 环境科学 遥感 地理 人工智能 物理 声学 地图学 业务 电信 营销 频道(广播) 多输入多输出
作者
Yucong Meng,Zhiwei Yang,Yonghong Shi,Zhijian Song
出处
期刊:Proceedings of the ... AAAI Conference on Artificial Intelligence [Association for the Advancement of Artificial Intelligence (AAAI)]
卷期号:39 (6): 6135-6143
标识
DOI:10.1609/aaai.v39i6.32656
摘要

The accelerated MRI reconstruction process presents a challenging ill-posed inverse problem due to the extensive under-sampling in k-space. Recently, Vision Transformers (ViTs) have become the mainstream for this task, demonstrating substantial performance improvements. However, there are still three significant issues remain unaddressed: (1) ViTs struggle to capture high-frequency components of images, limiting their ability to detect local textures and edge information, thereby impeding MRI restoration; (2) Previous methods calculate multi-head self-attention (MSA) among both related and unrelated tokens in content, introducing noise and significantly increasing computational burden; (3) The naive feed-forward network in ViTs cannot model the multi-scale information that is important for image restoration. In this paper, we propose FPS-Former, a powerful ViT-based framework, to address these issues from the perspectives of frequency modulation, spatial purification, and scale diversification. Specifically, for issue (1), we introduce a frequency modulation attention module to enhance the self-attention map by adaptively re-calibrating the frequency information in a Laplacian pyramid. For issue (2), we customize a spatial purification attention module to capture interactions among closely related tokens, thereby reducing redundant or irrelevant feature representations. For issue (3), we propose an efficient feed-forward network based on a hybrid-scale fusion strategy. Comprehensive experiments conducted on three public datasets show that our FPS-Former outperforms state-of-the-art methods while requiring lower computational costs.

科研通智能强力驱动
Strongly Powered by AbleSci AI
更新
PDF的下载单位、IP信息已删除 (2025-6-4)

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
刚刚
赘婿应助顺利寄文采纳,获得10
1秒前
Owen应助meinvqianer1219采纳,获得10
2秒前
4秒前
简单茗发布了新的文献求助10
4秒前
量子星尘发布了新的文献求助10
5秒前
5秒前
zzzzzzzzzzzz完成签到,获得积分10
6秒前
静汉发布了新的文献求助10
6秒前
青栀发布了新的文献求助10
10秒前
少川完成签到 ,获得积分10
11秒前
青糯完成签到 ,获得积分10
13秒前
静汉完成签到,获得积分10
14秒前
剪云者完成签到 ,获得积分10
19秒前
21秒前
21秒前
wenhao完成签到 ,获得积分10
22秒前
飞逝的快乐时光完成签到 ,获得积分10
22秒前
蟹蟹完成签到 ,获得积分10
23秒前
24秒前
25秒前
张劳西发布了新的文献求助10
27秒前
29秒前
tomoe发布了新的文献求助10
29秒前
30秒前
33秒前
科研通AI5应助科研通管家采纳,获得10
33秒前
大模型应助鱼仔采纳,获得10
33秒前
彭于晏应助科研通管家采纳,获得10
33秒前
科研通AI2S应助科研通管家采纳,获得10
34秒前
充电宝应助科研通管家采纳,获得10
34秒前
小二郎应助科研通管家采纳,获得30
34秒前
七慕凉应助科研通管家采纳,获得10
34秒前
34秒前
34秒前
猪猪hero应助科研通管家采纳,获得10
34秒前
猪猪hero应助科研通管家采纳,获得10
34秒前
彭于晏应助小颖采纳,获得10
35秒前
咻咻发布了新的文献求助20
35秒前
往事吴痕完成签到 ,获得积分10
36秒前
高分求助中
The Oxford Encyclopedia of the History of Modern Psychology 2000
Chinesen in Europa – Europäer in China: Journalisten, Spione, Studenten 1200
Deutsche in China 1920-1950 1200
Applied Survey Data Analysis (第三版, 2025) 850
Mineral Deposits of Africa (1907-2023): Foundation for Future Exploration 800
 Introduction to Comparative Public Administration Administrative Systems and Reforms in Europe, Third Edition 3rd edition 590
Learning to Listen, Listening to Learn 570
热门求助领域 (近24小时)
化学 材料科学 医学 生物 工程类 有机化学 物理 生物化学 纳米技术 计算机科学 化学工程 内科学 复合材料 物理化学 电极 遗传学 量子力学 基因 冶金 催化作用
热门帖子
关注 科研通微信公众号,转发送积分 3881415
求助须知:如何正确求助?哪些是违规求助? 3423887
关于积分的说明 10736313
捐赠科研通 3148707
什么是DOI,文献DOI怎么找? 1737444
邀请新用户注册赠送积分活动 838811
科研通“疑难数据库(出版商)”最低求助积分说明 784107