Deep Learning Based Approach: An Efficient Anomaly Detecting from Various Medical Signals and Medical Images Using CNNs

Abstract In the Artificial Intelligence (AI) field, the deep learning is considered a method falls in the wide machine learning algorithms family based on the learning principle. The known traditional and Conventional Neural Networks (CNNs) have been utilized in the pattern recognition techniques based on the deep learning concepts from different images. Due to the importance of the Anomaly Detection (AD) in automatic diagnosis, it is essential and vital point in the image and medical signal processing. In this paper, the AD has been tested and evaluated using the signals of the medical EEG employing the spectrogram and medical corneal images. The deep learning based on the CNN models are employed in the processes of training and testing, each image input passes through a convolution layers series and Kernels filters. For the classification, the pooling and Fully Connected (FC) layers have been utilized for this purpose. In this research paper, the experiments computer simulation have been presented, its results reveal that the success and superiority of the presented proposed techniques in the automated medical diagnosis.