作者: Shengran Su , Zhenghui Hu , Qiang Lin , William Kongto Hau , Zhifan Gao
DOI: 10.1016/J.COMPMEDIMAG.2016.11.003
关键词: Intravascular ultrasound 、 Softmax function 、 Artificial intelligence 、 Computer vision 、 Computer science 、 Lumen (anatomy) 、 Active contour model 、 Imaging technique 、 Feature learning 、 Artificial neural network
摘要: Intravascular ultrasound (IVUS) has been well recognized as one powerful imaging technique to evaluate the stenosis inside coronary arteries. The detection of lumen border and media-adventitia (MA) in IVUS images is key procedure determine plaque burden arteries, but this could be burdensome doctor because large volume images. In paper, we use artificial neural network (ANN) method feature learning algorithm for MA borders Two types information including spatial, neighboring features were used input data ANN method, then different vascular layers distinguished accordingly through two sparse auto-encoders softmax classifier. Another was optimize result first network. end, active contour model applied smooth detected by method. performance our approach compared with manual drawing performed experts on 461 from four subjects. Results showed that had a high correlation good agreement results. error close between groups result. All these results indicated proposed efficiently accurately handle