An Efficient and Optimized Convolution Neural Network for Covid and Lung Disease Detection

• Published in: 2023 8th International Conference on Communication and Electronics Systems (ICCES) pp. 735 - 740
• Main Authors: Agarwal, Mohit, Kaliyar, Rohit Kr, Gupta, Suneet Kr
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2023
• Subjects: Acceleration; Compression; Computational modeling; Convolution Neural Network; COVID-19; Image coding; Liver; Liver tumour; Medical diagnosis; Robots; Support vector machines
• DOI: 10.1109/ICCES57224.2023.10192708

Medical diagnosis has been widely enhanced by the deep learning methods using medical images such as X-rays, CT scans and MRI scans. The physical diagnosis by viewing the images can vary from one doctor to another. The deep learning based methods are found to produce more accurate results. This article proposes usage of transfer learning based pre-trained models such as VGG19, MobileNet, AlexNet, etc. Several traditional machine learning methods such as Logistic Regression, k-Nearest Neighbours (k-NN), Decision Trees (DT), Random Forest (RF), Support Vector Machine (SVM), Linear Discriminant Analysis (LDA), Naive Bayes have also been used to show different computer based methods for medical diagnosis. With the advent of robot based devices in various medical fields a need is created to deploy these models on low memory devices. Hence the pre-trained models which need more than 100 MBs space are compressed using Differential Evolution algorithm to reduce the space need to few KBs with similar accuracy.