Supervised machine learning for coronary artery lumen segmentation in intravascular ultrasound images

Abstract

Intravascular ultrasound (IVUS) has been widely used to capture cross sectional lumen frames of inner wall of coronary arteries. This kind of medical imaging modalities is capable of providing detailed and significant information of lumen contour shape, which is very important for clinical diagnosis and analysis of cardiovascular diseases. Numerous learning based techniques have recently become very popular for coronary artery segmentation due to their impressive results. In this work, a supervised machine learning method for coronary artery lumen segmentation with high accuracy and minimal user interaction is designed. The fully discriminative lumen segmentation method jointly learning a classifier the weak learners rely on and the features of the classifier is developed. Additionally, the theoretical supports of the Gradient Boosting framework used in this work and its quadratic approximation are presented. The proposed algorithm is tested on the public datasets of boundary detection of lumen in IVUS challenge held in MICCAI 2011 and achieves a higher average Jaccard similarity of 96.8% and a lower mean error distance of 0.55 (in Cartesian coordinates), which shows higher accuracy compared to the existing learning based methods. Moreover, three real patient IVUS datasets are used to evaluate the performance of the proposed coronary artery lumen segmentation algorithm, which is shown to achieve lower percent error of lumen area of 1.861% ± 0.965%, 1.968% ± 0.864%, and 1.671% ± 0.584%, respectively, compared to the manually measured lumen area (ground truth). The proposed lumen segmentation method is found to be superior to the latest learning based segmentation techniques. Given the efficiency and robustness, our method has great potential in IVUS images processing and coronary artery segmentation and quantification.Novelty StatementThe main contributions are summarized in the following aspects: A detailed review of related work about learning based coronary artery lumen segmentation in intravascular ultrasound images is presented. A fully discriminative lumen segmentation method jointly learning a classifier our weak learners rely on and the features of the classifier is developed. The theoretical supports of the Gradient Boosting framework and its quadratic approximation used in this work are presented.

RobUSt–An Autonomous Robotic Ultrasound System for Medical Imaging

“…[29], [30]. In this study we demonstrate that the roboticaided reconstruction of US volumes and the utilization of unsupervised learning method K-means enable the robust and automatic segmentation of sub-dermal tissues.…”

Deep Learning Paradigm and Its Bias for Coronary Artery Wall Segmentation in Intravascular Ultrasound Scans: A Closer Look

“…For the segmentation of the coronary walls, various AIbased techniques, such as ML-based or DL-based, have been applied. The ML-based method includes XGBoost [79,107,108], k-means [43], hidden Markov random field (HMRF) [43,109,110], support vector machine (SVM) [65,82], random forest (RF) [65,82], fuzzy c-means (FCM) [43,89], Pix2Pix model [74], ellipse-fitting algorithm [28], Lucky-Richardson algorithm [84], and gradient boosting [85]. The DL-based method includes generative adversarial network (GAN) [74], convolutional neural network (CNN) [78,81,95], bidirectional gated recurrent unit (Bi-GRU) [74], efficient net [75], DeepLabV3 [80], location-adaptive threshold method (LATM) [111], scan-adaptive threshold method (SATM) [111], and fully convolutional neural network (FCNN) [87].…”

“…Conventional method: (i) Otsu thresholding [90], (ii) Fuzzy method [87,89], (iii) Parametric deformable model [92], (iv) Geometric deformable model [92], (v) Gradient vector flow (GVF) [94], (vi) K-means [43], (vii) Lucky Richard algorithm [84], (viii) Ellipse fitting algorithm [28]. Machine Learning: (i) SVM [65,82], (ii) XGBoost [79,107,108], (iii) RF [65,82], (iv) Gradient boosting [85], (v) HMRF [43,109,110]. Deep Learning: (i) CNN [78,81,95], (ii) FCNN [87], (iii) Efficient-net [75], (iv) DeepLabV3 [80], (v) GAN [74], (vi) Pix-2-pix model [74], (vii) Bi-GRU [74], (viii) LSTM [97] (ix) LATM [111] (x) SATM [111].…”

“…The three types of bias employed in the VD process are shown in Figure 19a,c, including low bias (Figure 19a), moderate bias (Figure 19b), and high bias (Figure 19c). The number of studies in low bias (out of 28 studies) for RBM, RBA, and RBS was three (10%) [48,58,62], eight (28%) [45,49,50,57,60,61,64,69], and four (14%) [24,48,65,67] respectively, whereas the number of studies under low bias (out of 60 studies) for RBM, RBA, and RBS was seven (12%) [58,75,79,81,82,86,87], twelve (20%) [28,65,77,84,85,[89][90][91][92][93][94]111], and seven (12%) [24,48,55,58,59,62,67], respectively. Out of 28 and 60 studies, respectively, 2 [48,…”

“…The studies that fell under the intersection of (RBA, RBM, RBS), (RBA, RBM), and (RBM, RBS) was two, five, and six, respectively, whereas no common studies were found under the intersection of (RBA, RBS). Moreover, the number of studies under high bias (out of 60 studies) for RBM, RBA, and RBS was 44 (73%) [24,28,[44][45][46][47][48][49][50][51][52][53][54][55][56][57][59][60][61][62][64][65][66][67][68][69][70][71][72][73][74]77,84,85,[88][89][90][91][92][93][94]97,100,111], 28 (46%) [24,43,[45][46]...…”