A hybrid of deep and textural features to differentiate glomerulosclerosis and minimal change disease from glomerulus biopsy images

• Published in: Biomedical signal processing and control Vol. 70; p. 103020
• Main Authors: Santos, Justino Duarte, Veras, Rodrigo de M.S., Silva, Romuere R.V., Aldeman, Nayze L.S., Araújo, Flávio H.D., Duarte, Angelo A., Tavares, João Manuel R.S.
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01-09-2021
• Subjects: Deep learning; Feature extraction; Feature selection; Image analysis; Image classification; Deep learning; Feature extraction; Feature selection; Image analysis; Image classification
• DOI: 10.1016/j.bspc.2021.103020
• EISSN: 1746-8108
• ISSN: 1746-8094

The minimal change disease (MCD) and glomerulosclerosis (GS) are two common kidney diseases. Unless adequately treated, these diseases leads to chronic kidney diseases. Accurate differentiation of these two diseases is of paramount importance as their methods of treatment and prognoses are different. Thus, this article propose a method capable of differentiating MCD from GS in glomerulus biopsies images based on a new hybrid deep and texture feature space. We conducted an extensive study to determine the best set of features for image representation. Our feature extraction methodology, which includes Haraliks and geostatistics texture descriptors and pre-trained CNNs, resulted in 13,476 characteristics. We then used mutual information to order the elements by importance and select the best set for differentiating MCD from GS using the random forest classifier. The proposed method achieved an accuracy of 90.3% and a Kappa index of 80.5%. Representation of glomerulus biopsy images with a hybrid of deep and textural features facilitates the accurate differentiation of GS and MCD.
•An automated method to analyze glomerulus biopsy images.•Accurately differentiates Glomerulosclerosis & Minimal Change Disease.•Advantages of using deep features and textural features are combined.•Highly accurate.•An experiment set combining 13,476 features, 2 filtering methods, and 2 classifiers.