Texture Feature Extraction from Intracoronary OCT Images and Atherosclerosis Detection using Deep Neural Network

A coronary atherosclerotic plaque's morphological structure and tissue composition evaluate its stability, this could be investigated using intracoronary optical coherence
tomography (OCT) imaging. The objective of this research is to extract intensity features from OCT images utilising Histograms of Oriented Gradients (HOG) as well as Local
Binary Patterns (LBP) bag-of-words (BOW). The approach is concentrated on Twopath Convolution Neural Network(CNN),an unique CNN architecture. This proposed technique is reliable and durable in the detection of atherosclerosis OCT imaging, and per the evaluation. The accuracy of the Twopath CNN architecture is significantly higher than that of conventional CNN methodologies as well as machine learning approaches. This method seems to have a better efficiency of 98.5 %, suggesting that it could be a suitable diagnostic tool for detecting atherosclerosis. Keywords-CNN, HOG, atherosclerosis