Resumen

BERROUANE, Nadir; BENYETTOU, Mohammed  y  IBTISSAM, Benchennane. Deep Learning and Feature Extraction for Covid 19 Diagnosis. Comp. y Sist. [online]. 2022, vol.26, n.2, pp.909-920.  Epub 10-Mar-2023. ISSN 2007-9737.  https://doi.org/10.13053/cys-26-2-4268.

Recently, medical images analysis is becoming the center of interest in the medical field, with the helpful opportunities offered by artificial intelligence, especially deep learning techniques. Computers are becoming more and more capable of learning how to be diagnosing certain medical pathologies and diseases. In this domain, deep learning is a major choice, more precisely Convolutional Neural Networks (CNN) due to its powerful performance with images classification. In this paper, a new approach is proposed which is about using feature extraction from images and deep learning algorithms to avoid the issue of the necessity of a large dataset. This work aims to improve the diagnostic of the Covid 19 virus in X-ray images, by extracting the features and applying the deep learning algorithm. This approach is composed of two main phases. The first one is based on feature extraction from images using feature extraction algorithms: Pyramid Histogram of Gradient (PHOG), Fourier, Gabor, and Discrete Cosine Transform (DCT). The second phase is based on using the last layers of CNN of deep learning for the classification problem. The experimentation of our approach is demonstrated by utilizing chest X-ray images obtained by PyImageSearch. Analysis of results shows that the proposed approach provides a satisfactory result. Our approach could be so beneficial in the future that it can be used to solve real-life problems even though insufficient data especially in urgent cases where there is not enough time to collect the data.

Palabras llave : Classification; feature extraction; deep learning.

        · texto en Inglés     · Inglés ( pdf )