Scholarlink

Diagnosing cardiovascular diseases is challenging and time-consuming, yet it is a critical global health concern. Algorithm analysis of ECG images provides significant information; however, the pure essence of deep learning is often met with distrust by healthcare professionals, thus limiting its adoption. The objective of this paper is to maintain the integrity of AI based cardiac interpretation by employing XAI strategies such as Grad-CAM and SHAP. Considering the progress made in deep learning architectures which include CNNs and LSTMs, as well as the limitations that have been identified, such as the lack of large datasets and appropriate evaluation techniques – the study offers a comprehensive novel framework seeking to foster the comprehension and trustworthiness of AI tools. Such a method benefits patients by clarifying AI tools’ features along with the models’ predictions to medical practitioners and aids their broader acceptance.

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

Unveiling the Black Box: Enhancing Acceptance and Trust in ECG Image-Based Cardiac Diagnosis, Employing Explainable AI Techniques

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • R. Shyama,
  • M. Mythily

摘要

Diagnosing cardiovascular diseases is challenging and time-consuming, yet it is a critical global health concern. Algorithm analysis of ECG images provides significant information; however, the pure essence of deep learning is often met with distrust by healthcare professionals, thus limiting its adoption. The objective of this paper is to maintain the integrity of AI based cardiac interpretation by employing XAI strategies such as Grad-CAM and SHAP. Considering the progress made in deep learning architectures which include CNNs and LSTMs, as well as the limitations that have been identified, such as the lack of large datasets and appropriate evaluation techniques – the study offers a comprehensive novel framework seeking to foster the comprehension and trustworthiness of AI tools. Such a method benefits patients by clarifying AI tools’ features along with the models’ predictions to medical practitioners and aids their broader acceptance.