Scholarlink

Globally, cardiovascular diseases are the primary cause of death. It’s critical to have a non-invasive, efficient early detection tool. This paper introduces a novel approach for cardiac risk assessment using retinal fundus. A CNN-Transformer-Ensemble model is being suggested that utilizes the strengths of both methods. To enable the model to recognize patterns from one layer to the next, our methodology starts with the extraction of data from the retinal fundus using a CNN. The ensemble model uses Transformer architecture to comprehend long-range dependencies between the layers. The results obtained demonstrate an exceptional performance with the model achieving an overall accuracy of 97%.

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

Identification of Cardiac Risk from Retinal Fundus Image Using Hybrid CNN-Transformer-Ensemble Model

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Manjula Sri Rayudu,
  • Naga Dheeraj Gopu,
  • Venkanna Chenagoni,
  • Rama Valupadasu

摘要

Globally, cardiovascular diseases are the primary cause of death. It’s critical to have a non-invasive, efficient early detection tool. This paper introduces a novel approach for cardiac risk assessment using retinal fundus. A CNN-Transformer-Ensemble model is being suggested that utilizes the strengths of both methods. To enable the model to recognize patterns from one layer to the next, our methodology starts with the extraction of data from the retinal fundus using a CNN. The ensemble model uses Transformer architecture to comprehend long-range dependencies between the layers. The results obtained demonstrate an exceptional performance with the model achieving an overall accuracy of 97%.