DeepHeart: Leveraging Multitask Convolutional Neural Networks for Cardiovascular Risk Assessment

Heart disease is one of the leading causes of illness and death today. It poses a significant health concern in our everyday lives, and treating it remains quite challenging. Heart disease cases have been rising quickly lately, so it's critical to anticipate these conditions. It is challenging to identify the conditions, and because it involves sensitive information, it must be done correctly. In the existing system various machine learning algorithms like Support Vector Machine, Random Forest, Adaboost were used but it is inefficient for huge datasets and it is time consuming. To overcome this and accurately predict the heart disease, this study has proposed a novel system to compare a 2D convolutional neural network with a Multi-Task Cascaded Convolutional Neural Network. This study aims to accurately identify potential heart disease indicators by using the MTCNN for robust feature detection and alignment in medical images, followed by a CNN trained on these aligned images. The proposed approach provides a non-invasive and effective means of diagnosis by leveraging deep learning's capacity to automatically learn discriminative features from raw data. The proposed system has demonstrated effective prediction of heart disease with an accuracy of 99.4 % through extensive experimentation and validation. This suggests promising opportunities for early detection and intervention in cardiovascular health.