 This paper presents a novel approach to detect coronary artery stenosis using computer vision techniques. The proposed methodology consists of two stages. Firstly, it uses UNET, RESUNET A, UNET++, RESUNET A and UNET++ models for automatic coronary artery segmentation. Secondly, it employs DenseNet 201, EfficientNet B0, MobileNet V2, ResNet 101 and section models for coronary artery classification. The results show that UNET model achieved the best performance with a dice score of 0.8467 and a jacket index of 0.7454. Additionally, DenseNet 201 outperformed all other pre-trained models with an accuracy of 0.900, with a specificity of 0.9833, positive predictive value of 0.9556, Cohen's Kappa of 0.7746 and AUC of 0.9694. This article was authored by Serif Kaba, Hussain Hasey, Ali Aisin and others. We are article.tv, links in the description below.