期刊名称：Indian Journal of Computer Science and Engineering
印刷版ISSN：2231-3850
电子版ISSN：0976-5166
出版年度：2021
卷号：12
期号：6
页码：1611-1620
DOI：10.21817/indjcse/2021/v12i6/211206028
语种：English
出版社：Engg Journals Publications
摘要：Image classification task plays a vital role in the Computer-Aided Diagnosis system for diagnosis of prostate cancer. Transrectal Ultrasonography (TRUS) imaging techniques are mostly used by physicians to distinguish the prostate region from tissues around it and the same used for abnormality detection. The removal of irrelevant regions from the TRUS image is very rigid since TRUS images contain speckle noise, low dissimilarity, the fuzzy region between object and background, and also irregular form and size. To resolve the problem, the images are preprocessed by Ant Colony Optimization (ACO) method, and the relevant features are segmented by using Ant Colony Optimization- Boundary Complete Recurrent Neural Network (ACO-BCRNN). Then, VGG-19 transfer learning techniques are used to extract the Region of Interest (RoI) features. The relevant features are selected by using hybridization of the Salp Swarm Optimization Algorithm and Grasshopper Optimization Algorithm (SSOAGOA). In this article, Convolutional Neural Network with Bi-directional Long Short Term Memory (CNN-BiLSTM) is proposed for the classification task to distinguish the normal from abnormal image of prostate gland. The performance of the proposed method is assessed by using the classifiers such as Support Vector Machine (SVM), Grid Search (GS) and Extreme Machine Learning (ELM). The comparative analysis of proposed method is performed with CNN and CNN-LSTM. The result of the proposed method shows a significant gain in accuracy than other methods.
