Title: Automotive knee osteoarthritis severity assessment by adaptive multiscale deep learning techniques

Authors: Sriramulu Devarapaga; Rajesh Thumma

Addresses: Anurag University, Hyderabad, Telangana 500088, India ' Department of Electronics and Communication Engineering, Anurag University, Hyderabad, Telangana 500088, India

Abstract: Knee OsteoArthritis (KOA) is mostly occurring in the old aged people. Manual diagnosis involves X-ray for classifying KOA disorder whereas the physician's expertise requires more time and also it prone to errors. The consideration of a larger dataset in the training process causes a variety of related problems like processing the data consumes a huge time and also it cause overfitting issues. Hence, an enhanced deep learning model is introduced to resolve aforementioned challenges. Initially, the acquired images are considered further given into Adaptive Multiscale Residual DenseNet with Gated Recurrent Unit (AMRD-GRU). If the KOA is present, then the severity of the KOA is determined by the same AMRD-GRU technique. Here, the performance of the severity assessment is boosted to tune parameters from the AMRD-GRU using the opposition lyrebird optimisation algorithm (OLBO). The designed framework is validated to show the efficiency of the model using diverse measures.

Keywords: knee osteoarthritis; disease classification; adaptive multiscale residual denseNet with gated recurrent unit; severity assessment; OLBO; opposition lyrebird optimisation algorithm.

DOI: 10.1504/IJSISE.2024.143822

International Journal of Signal and Imaging Systems Engineering, 2024 Vol.13 No.3, pp.186 - 202

Received: 06 Apr 2024
Accepted: 11 Sep 2024
Published online: 08 Jan 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article