Novel superpixel‐based algorithm for segmenting lung images via convolutional neural network and random forest

Abstract

Accurately segmenting lungs from CT images is a fundamental step for quantitative analysis of lung diseases. However, it is still a challenging task because of some interferential factors, such as juxta-pleural nodules, pulmonary inflammation, as well as individual anatomical varieties. In this study, with the combination of a superpixel approach and a hybrid model composed of convolutional neural network and random forest (CNN-RF), the authors propose a novel algorithm to segment lungs from CT images in an automatic and accurate fashion. The authors' lung segmentation covers three main stages: image preprocessing, lung segmenting and segmentation refining. A lung CT image denoised with a fractional-order grey similarity approach is first segmented to a set of superpixels, and the CNN-RF model is then employed to classify the superpixels and identify lungs from the CT image. The segmentation result is further refined by separating the left and right lungs, eliminating trachea, and correcting lung contours. Experiments show that their algorithm can generate more accurate lung segmentation results with 94.98% Jaccard's index and 97.99% Dice similarity coefficient, compared with ground truths, and it achieved better results compared with several feature-based machine learning techniques and current methods on lung segmentation.

A Systematic Review of Automated Segmentation Methods and Public Datasets for the Lung and its Lobes and Findings on Computed Tomography Images

“…They proposed the following classification: neighboring anatomy-guided [13]; thresholding [12,14]; region-based [2,15]; shape or model-based [16]; machine learning [17] and hybrids [18]. Hybrids include methods where it is not clear to which category of method it fits, usually because they use a combination of multiple types of approaches [18][19][20][21][22][23][24][25][26][27][28][29][30][31][32]. We concluded that while this classification still applies, we noticed the necessity to separate machine learning into two categories: traditional machine learning and deep learning [33], due to the recent explosion of deep learning methods.…”

Extreme learning machine algorithm‐based model for lung cancer classification from histopathological real‐time images

“…This classifier includes three phases: preprocessing, segmentation, and optimization. 12 Firstly, a denoised lung scan was segmented through a set of superpixels, and then a classification algorithm was employed to detect lung cancer. Xie et al proposed a classifier using deep convolutional neural networks (DCNN) for identifying exact spots of lung cancer in CT scans rapidly.…”

“…The minimal cut value is calculated from Equation ( 12) from the overall similarity calculation. The NP-hardness of the similarity calculation issue is conquered by solving the minimal cut value given by Equation (12).…”

“…Liu et al proposed an efficient classifier using the superpixels method with a hybrid classifier using CNN and RF algorithms. This classifier includes three phases: preprocessing, segmentation, and optimization 12 . Firstly, a denoised lung scan was segmented through a set of superpixels, and then a classification algorithm was employed to detect lung cancer.…”

Research on Management Efficiency and Dynamic Relationship in Intelligent Management of Tourism Engineering Based on Industry 4.0

“…The research results of Kai et al show that the “data outside” interaction model based on discrete dynamic modeling technology is better than the traditional interaction model in terms of motion data information acquisition ability [ 14 ]. In order to improve the acquisition efficiency of tourism engineering management data and the maximum utilization of storage space, Liu et al conducted various research studies and analyses on different tourism engineering management data and finally classified different data types into various types of data models to avoid storing duplicate data and increase the collection efficiency of tourism data [ 15 ]. Goodarzian et al proved through experiments that the discrete data mining model established can realize the static segmentation of dynamic data and solve the problem that it is difficult to obtain dynamic data in management, but there is a problem of small application range [ 16 ].…”