MedicalCLIP: Anomaly-Detection Domain Generalization with Asymmetric Constraints
-
Published:2024-05-16
Issue:5
Volume:14
Page:590
-
ISSN:2218-273X
-
Container-title:Biomolecules
-
language:en
-
Short-container-title:Biomolecules
Author:
Hua Liujie1ORCID, Luo Yueyi2, Qi Qianqian3, Long Jun3
Affiliation:
1. School of Computer Science and Engineering, Central South University, Changsha 410083, China 2. School of Mathematics and Statistics, Central South University, Changsha 410083, China 3. Big Data Institute, Central South University, Changsha 410083, China
Abstract
Medical data have unique specificity and professionalism, requiring substantial domain expertise for their annotation. Precise data annotation is essential for anomaly-detection tasks, making the training process complex. Domain generalization (DG) is an important approach to enhancing medical image anomaly detection (AD). This paper introduces a novel multimodal anomaly-detection framework called MedicalCLIP. MedicalCLIP utilizes multimodal data in anomaly-detection tasks and establishes irregular constraints within modalities for images and text. The key to MedicalCLIP lies in learning intramodal detailed representations, which are combined with text semantic-guided cross-modal contrastive learning, allowing the model to focus on semantic information while capturing more detailed information, thus achieving more fine-grained anomaly detection. MedicalCLIP relies on GPT prompts to generate text, reducing the demand for professional descriptions of medical data. Text construction for medical data helps to improve the generalization ability of multimodal models for anomaly-detection tasks. Additionally, during the text–image contrast-enhancement process, the model’s ability to select and extract information from image data is improved. Through hierarchical contrastive loss, fine-grained representations are achieved in the image-representation process. MedicalCLIP has been validated on various medical datasets, showing commendable domain generalization performance in medical-data anomaly detection. Improvements were observed in both anomaly classification and segmentation metrics. In the anomaly classification (AC) task involving brain data, the method demonstrated a 2.81 enhancement in performance over the best existing approach.
Funder
Systemically significant projects of China National Railway Group Co., I LTD National Natural Science Foundation of China National 745 Natural Science Foundation of China Hunan Provincial Natural Science Foundation of China High-Performance Computing Center of Central South University
Reference47 articles.
1. Liu, Z., Zhou, Y., Xu, Y., and Wang, Z. (2023, January 17–24). Simplenet: A simple network for image anomaly detection and localization. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, BC, Canada. 2. Roth, K., Pemula, L., Zepeda, J., Schölkopf, B., Brox, T., and Gehler, P. (2022, January 18–24). Towards total recall in industrial anomaly detection. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, LA, USA. 3. Jiang, H., Dang, Z., Wei, Z., Xie, J., Yang, J., and Salzmann, M. (2023, January 17–24). Robust Outlier Rejection for 3D Registration with Variational Bayes. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, BC, Canada. 4. DoTA: Unsupervised detection of traffic anomaly in driving videos;Yao;IEEE Trans. Pattern Anal. Mach. Intell.,2022 5. Few-shot domain-adaptive anomaly detection for cross-site brain images;Su;IEEE Trans. Pattern Anal. Mach. Intell.,2021
|
|