Downloads: 3 | Views: 69 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1
Research Paper | Computer Science and Information Technology | India | Volume 12 Issue 10, October 2023
Bioinspired Optimization with Deep Learning Driven Cardiovascular Diseases on Retinal Fundus Images
R. Ramesh [36] | Dr. S. Sathiamoorthy
Abstract: Earlier diagnosis of cardiovascular diseases (CVDs) is important for timely intervention and improved survival rates. Retinal fundus images are developed as a precious diagnostic resource owing to their non-invasive nature and capability to reveal vascular abnormalities representative of CVDs. This study designs an automated Cardiovascular diseases using Seagull Optimization Algorithm with Deep Learning (ACVD-SOADL) method on retinal fundus images. This study introduces a wide-ranging approach for CVD diagnosis by leveraging DL approaches, particularly employing MobileNet as a feature extractor, Back propagation Neural Network (BPNN) for classification, and Seagull Optimization Algorithm (SOA) for parameter tuning. MobileNet, a lightweight DL model, is utilized for extracting meaningful features from these images. SOA is an innovative optimizer algorithm stimulated by the foraging behavior of seagulls, which can be employed to fine-tune the model's hyperparameters, improving its performance. The dataset exploited in this study includes a various types of retinal fundus images, allowing generalizable and robust training method. The BPNN classification method is trained to differentiate among normal and CVD-affected retinal images depends on the extracted features. By employing an iterative process, SOA enhances the hyper parameters of the BPNN, confirming that the model attains its highest potential accuracy. For demonstrating the improvised performance of the ACVD-SOADL approach, an extensive simulated value could be accomplished and the comparison analysis assured the excellence of the ACVD-SOADL model.
Keywords: Fundus images, Deep learning, Cardiovascular diseases, Seagull optimization algorithm, Artificial intelligence
Edition: Volume 12 Issue 10, October 2023,
Pages: 1869 - 1876