FuzzyNet: Medical Image Classification based on GLCM Texture Feature

In recent study has found that high-precision monitoring, big data, and medical diagnosis continue to be hampered by transit delays. We designed and implemented an unique fuzzy logic-based method to help with this problem. Currently, a high-accuracy automated approach for detecting anomalies in X-ray images is being developed. Pre-processing image technologies are used to improve the quality of medical pictures in order to achieve high accuracy while utilizing a small number of system resources. Noise reduction and contrast enhancement are two procedures engaged in image pre-processing that assist to the delivery of a rapid anomaly detecting system. We suggested a Fuzzy Net model Classification Based on GLCM Feature and particle swam optimization technique in this study. This method categorizes the X-ray pictures. The musculoskeletal radiographs (MURA) dataset is classified as normal or abnormal. We also examined at performance indicators and loss over epochs, which we compared using a confusion matrix and displayed plots of the model’s learned membership functions.

I. Introduction

In today’s society, these models have undergone significant training and have attained extraordinary levels of accuracy. However, a feature used for classification cannot be a perfect match to a crisp number, implying that there is some fuzziness in the data [1]. Due to its great dimensionality and low cost, the GLCM (Gray Level Co-occurrence Matrix) texture characteristics are frequently used in picture classification problems. The second-order statistical information about the grey levels between neighboring pixels in the picture is provided by the grey level GLCM [2]. With the aid of the automated system, several advanced features are employed to identify objects that humans are unable to recognize on their own. When it comes to machine intelligence, the new coming age of intelligent machines offers radiologists tremendous assistance in their clinical processes.