https://scholars.lib.ntu.edu.tw/handle/123456789/607100
標題: | Classification of squamous cell carcinoma from FF-OCT images: Data selection and progressive model construction | 作者: | Ho C.-J Calderon-Delgado M Lin M.-Y Tjiu J.-W Huang S.-L HOMER H. CHEN |
關鍵字: | Convolutional neural network;Deep learning;Optical coherence tomography;Regularization;Squamous cell carcinoma;Training strategy;Convolution;Convolutional neural networks;Data reduction;Deep neural networks;Image classification;Image enhancement;Large dataset;Statistical tests;Data Selection;Full-field optical coherence tomographies;Image data;Model construction;Progressive models;Regularisation;Optical tomography;Article;cancer classification;cancer diagnosis;classification algorithm;classifier;comparative study;computer assisted diagnosis;convolutional neural network;deep learning;diagnostic accuracy;feature selection;full field optical coherence tomography;human;optical coherence tomography;residual neural network;squamous cell skin carcinoma;three-dimensional imaging;animal;diagnostic imaging;mouse;squamous cell carcinoma;Animals;Carcinoma, Squamous Cell;Mice;Neural Networks, Computer;Tomography, Optical Coherence | 公開日期: | 2021 | 卷: | 93 | 來源出版物: | Computerized Medical Imaging and Graphics | 摘要: | We investigate the speed and performance of squamous cell carcinoma (SCC) classification from full-field optical coherence tomography (FF-OCT) images based on the convolutional neural network (CNN). Due to the unique characteristics of SCC features, the high variety of CNN, and the high volume of our 3D FF-OCT dataset, progressive model construction is a time-consuming process. To address the issue, we develop a training strategy for data selection that makes model training 16 times faster by exploiting the dependency between images and the knowledge of SCC feature distribution. The speedup makes progressive model construction computationally feasible. Our approach further refines the regularization, channel attention, and optimization mechanism of SCC classifier and improves the accuracy of SCC classification to 87.12% at the image level and 90.10% at the tomogram level. The results are obtained by testing the proposed approach on an FF-OCT dataset with over one million mouse skin images. ? 2021 Elsevier Ltd |
URI: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85116513776&doi=10.1016%2fj.compmedimag.2021.101992&partnerID=40&md5=aefaac39b7ab847e21268075e1a05a2d https://scholars.lib.ntu.edu.tw/handle/123456789/607100 |
ISSN: | 08956111 | DOI: | 10.1016/j.compmedimag.2021.101992 |
顯示於: | 電機工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。