A Survey on Neural and Non-Neural Network Based Approaches to Classify Images and Signals

Abstract

The classification of signals and images using machine learning and artificial intelligence is a rapidly growing field with various applications across various industries. It is used in diverse areas of life, from medical science to security and transportation to entertainment. The ability to classify and analyze signals and images using ML and AI techniques allows for improved automation, decision-making, and predictions in many fields. This research examines the classification of signals and images using various neural and non-neural network-based algorithms. The focus is on the application of Convolutional Neural Networks (CNN) on image and signal datasets, specifically in the medical field. The classification of EEG signals is used to identify epileptic seizure disease, while food image datasets are used to classify seven different categories of food. Additionally, five pre-trained CNN models were applied to the food dataset using transfer learning techniques, with the VGG19 model achieving the highest accuracy of 94%. The classification of EEG signals using a publicly available dataset resulted in an accuracy of 98%. This study highlights the potential of machine learning in the analysis and classification of medical images and signals and the ability of CNNs to classify such data effectively.