Automated Chest X-Ray Image Classification using Manta Ray Optimization with Deep Learning Approach

Abstract

Medical image classification has played a key role in teaching tasks and clinical treatment. But the conventional technique has reached its maximum performance. Furthermore, more effort and time should be spent on selecting and extracting classification features by using them. The deep neural network is an emergent machine learning (ML) technique that has shown its potential for numerous classification tasks. Especially, the convolution neural network predominates with better outcomes on different image classification models. This article develops an Automated Chest X-Ray Image Classification using Manta Ray Optimization with Deep Learning Approach (MRFO-DLA). The presented MRFO-DLA technique majorly concentrates on the recognition and classification of diseases using CXRs. To attain this, the presented MRFO-DLA technique designs a bilateral filtering (BF) approach is applied for image preprocessing stage. Next, the MRFO-DLA model employs neural architectural search network (NASNet) for feature extraction. At the final stage, the MRFO with autoencoder (AE) model is exploited for CXR classification. To demonstrate the enhanced performance of the MRFO-DLA method, a series of simulations were performed and the outcomes were inspected in diverse aspects. The simulation outcomes ensured the enhancements of the MRFO-DLA approach compared to recent techniques interms of different measures.