Scholarlink

<p>Coronary Artery Disease (CAD) is the greatest prevalent cardiovascular disease (CVD) that frequently heart attacks. Lots of deaths and billions of greenbacks internationally occur annually due to invasive and expensive coronary artery disease detection through angiography. Deep Learning (DL) techniques are cited as affordable, fast, and non-invasive CAD prediction methods in the literature. However, outcomes vary significantly depending on datasets, traits, data collection spots, sample sizes, performance metrics, and applied DL approaches. AI(Artificial Intelligence) assisted CAD in medical imaging has made progress, but improvements are needed for precise disease diagnosis, despite basic disparities in progress. Intending to achieve maximum accuracy, this paper presents a novel DL framework for CAD. Initially, the input images undergo a pre-processing phase using an Anisotropic Diffusion Filter (ADF) for noise reduction and a Contrast Normalization (CN) to get normalized images. An Optimized Deep Neural Network (ODNN) remained built to forecast the illness. To attain maximum accurateness, a fitness-dependent improved seagull optimization algorithm (ISOA) technique is adopted to pick the optimal features and thereby increase prediction performance with an accuracy of 98.58%. The competence of the proposed ODNN-ISOA is evaluated with rigorous performance metrics and compared with SOTA models.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

RETRACTED ARTICLE: An automated coronary heart disease prediction system using optimized deep neural network framework

  • 期刊论文

  • 发表日期：

    2025年5月16日

查看全文
  • C. Venkata Narasimhulu

摘要

Coronary Artery Disease (CAD) is the greatest prevalent cardiovascular disease (CVD) that frequently heart attacks. Lots of deaths and billions of greenbacks internationally occur annually due to invasive and expensive coronary artery disease detection through angiography. Deep Learning (DL) techniques are cited as affordable, fast, and non-invasive CAD prediction methods in the literature. However, outcomes vary significantly depending on datasets, traits, data collection spots, sample sizes, performance metrics, and applied DL approaches. AI(Artificial Intelligence) assisted CAD in medical imaging has made progress, but improvements are needed for precise disease diagnosis, despite basic disparities in progress. Intending to achieve maximum accuracy, this paper presents a novel DL framework for CAD. Initially, the input images undergo a pre-processing phase using an Anisotropic Diffusion Filter (ADF) for noise reduction and a Contrast Normalization (CN) to get normalized images. An Optimized Deep Neural Network (ODNN) remained built to forecast the illness. To attain maximum accurateness, a fitness-dependent improved seagull optimization algorithm (ISOA) technique is adopted to pick the optimal features and thereby increase prediction performance with an accuracy of 98.58%. The competence of the proposed ODNN-ISOA is evaluated with rigorous performance metrics and compared with SOTA models.