干旱区生态环境知识资源中心(Arid): Adversarial image reconstruction learning framework for medical image retrieval

Arid

DOI	10.1007/s11760-021-02070-6
	Adversarial image reconstruction learning framework for medical image retrieval
	Pinapatruni, Rohini; Bindu, Chigarapalle Shoba
通讯作者	Pinapatruni, R (corresponding author)，JNTUA, Anatapuramu, India.
来源期刊	SIGNAL IMAGE AND VIDEO PROCESSING
ISSN	1863-1703
EISSN	1863-1711
出版年	2022-01
英文摘要	Due to the advancement in digital recording techniques, a large amount of data (images/video) is created in medical centers. Penetrating this indistinguishable data is a challenging task for many healthcare applications. Handling massive data with hand-crafted feature-based techniques is a difficult and time-consuming task. To solve this problem, a robust adversarial image reconstruction learning framework is proposed for medical image retrieval. The proposed approach consists of two stages viz feature extraction through adversarial image reconstruction and index matching followed by retrieval module to retrieve the similar images to that of input medical image. Initially, the adversarial image reconstruction network (AIR-Net) is proposed to encode the input medical image into set of features followed by the reconstruction of the input medical image from the encoded features. These encoded features give latent representation for robust reconstruction for the input image. Therefore, these encoded features are used in the index matching and retrieval module for medical image retrieval task. We use the self-attention mechanism in the proposed AIR-Net to suppress feature redundancy and enhance feature learning ability. The performance of the proposed framework is analyzed on benchmark medical image databases such as OASIS, ILD, VIA/ELCAP-CT for image retrieval task. To examine the robustness of the proposed framework over the existing state-of-the-art approaches, the retrieval accuracy in terms of average precision and recall is compared. From the experimental analysis, it is observed that the proposed approach outperforms the other existing approaches for medical image retrieval task
英文关键词	Self-attention block Adversarial Learning Index matching Medical image retrieval
类型	Article ; Early Access
语种	英语
收录类别	SCI-E
WOS记录号	WOS:000741619300001
WOS关键词	LOCAL TERNARY PATTERNS ; FEATURE DESCRIPTOR ; TEXTURE ; CLASSIFICATION ; MRI
WOS类目	Engineering, Electrical & Electronic ; Imaging Science & Photographic Technology
WOS研究方向	Engineering ; Imaging Science & Photographic Technology
资源类型	期刊论文
条目标识符	http://119.78.100.177/qdio/handle/2XILL650/377153
作者单位	[Pinapatruni, Rohini; Bindu, Chigarapalle Shoba] JNTUA, Anatapuramu, India
推荐引用方式 GB/T 7714	Pinapatruni, Rohini,Bindu, Chigarapalle Shoba. Adversarial image reconstruction learning framework for medical image retrieval[J],2022.
APA	Pinapatruni, Rohini,&Bindu, Chigarapalle Shoba.(2022).Adversarial image reconstruction learning framework for medical image retrieval.SIGNAL IMAGE AND VIDEO PROCESSING.
MLA	Pinapatruni, Rohini,et al."Adversarial image reconstruction learning framework for medical image retrieval".SIGNAL IMAGE AND VIDEO PROCESSING (2022).