Detection of Cardiovascular Diseases in ECG Images Using Image Processing and Deep Learning

Recent medical research on cardiovascular diseases (CVD) has affected people of younger age because of high stress in their personal lives, and CVD affects blood vessels. Finding the type of CVD using an ECG Heart signal dataset has faced challenges. To overcome this limitation, CVD was identified using a deep learning algorithm, specifically Inception v4, combined with an image processing technique-the first step in image preprocessing involved applying a Wiener filter to reduce the noise area. The second step is the Active Region Growing (ARG) segmentation approach, which differentiates the affected area and outer layer of the heart vessel image to evaluate the features. The third Sequential Random Elimination (SRE) categorizes the exact features based on the threshold values. Finally, Inception v4 classifies a specific CVD disease using improved performance analysis. The output evaluated the performance analysis in terms of specificity (97.23%) and accuracy (97.73%), indicating a reliable diagnosis using a simulated experiment.