3D skeletonization feature based computer-aided detection system for pulmonary nodules in CT datasets

Zhang, Weihang; Wang, Xue; Li, Xuanping; Chen, Junfeng

doi:10.1016/j.compbiomed.2017.11.008

Cited by 48 publications

(27 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, although many methods of lung nodule detection has been proposed [7][8][9] , it is still difficult to obtain satisfactory detection result due to the heterogeneity of lung nodules on CT images (as shown in Fig. 1).…”

Section: Introductionmentioning

confidence: 99%

A Two-Stage Convolutional Neural Networks for Lung Nodule Detection

Cao

Liu

et al. 2020

IEEE J. Biomed. Health Inform.

View full text Add to dashboard Cite

Early detection of lung cancer is an effective way to improve the survival rate of patients. It is a critical step to have accurate detection of lung nodules in computed tomography (CT) images for the diagnosis of lung cancer. However, due to the heterogeneity of the lung nodules and the complexity of the surrounding environment, robust nodule detection has been a challenging task. In this study, we propose a two-stage convolutional neural network (TSCNN) architecture for lung nodule detection. The CNN architecture in the first stage is based on the improved UNet segmentation network to establish an initial detection of lung nodules. Simultaneously, in order to obtain a high recall rate without introducing excessive false positive nodules, we propose a novel sampling strategy, and use the offline hard mining idea for training and prediction according to the proposed cascaded prediction method. The CNN architecture in the second stage is based on the proposed dual pooling structure, which is built into three 3D CNN classification networks for false positive reduction.Since the network training requires a significant amount of training data, we adopt a data augmentation method based on random mask. Furthermore, we have improved the generalization ability of the false positive reduction model by means of ensemble learning. The proposed method has been experimentally verified on the LUNA dataset.Experimental results show that the proposed TSCNN architecture can obtain competitive detection performance. Keywords: lung nodule detection; UNet; 3D CNN; ensemble learning;computer-aided diagnosis methods in the following aspects: 1) using the improved UNet segmentation model for lung nodule detection; 2) for the segmentation model training, we propose a new sampling strategy and an offline hard mining training approach; 3) we propose a cascade prediction method different from the traditional prediction method; 4) build three 3D CNN classification networks based on the dual pooling method; 5) design a data augmentation method based on random mask. MethodsThe lung nodule detection framework proposed in this paper is divided into two stages. The first stage: the detection of candidate nodules, which is based on the UNet architecture to achieve the detection of candidate nodules by segmenting suspicious nodules. The second stage: the reduction of false positive nodules, which is based on the 3DCNN architecture to eliminate false positive nodules through the integration of multiple models. The overall architecture of the proposed lung nodule detection method is shown in Fig. 2.

show abstract

Section: Introductionmentioning

confidence: 99%

A Two-Stage Convolutional Neural Networks for Lung Nodule Detection

Cao

Liu

et al. 2020

IEEE J. Biomed. Health Inform.

View full text Add to dashboard Cite

show abstract

“…Recently, Zhang et al [13] also developed a CAD for nodule detection that applied 3D skeleton features. For lung region segmentation, they used the global optimal active contour model, which can enclose both solitary and boundary attached nodules.…”

Section: Related Workmentioning

confidence: 99%

“…Generally, nodules are circular, while other lesions, especially vessels and fissures, are tubular in their structure. Based on this fact, we calculated three criteria, namely, surface area, eccentricity, and voxel remove rate (VRR) [13] in order to check the shape of each candidate object. Let be a labeled object of a roughly segmented lung structure, and its surface area can be calculated by…”

Section: Preliminary Screeningmentioning

confidence: 99%

“…Moreover, they are less sensitive to noises and blur boundaries because they neglect spatial information about the image. Unlike the intensity-based methods, region-based segmentation schemes [10][11][12][13] consider not only spatial information, but also the relationship between the pixels. To segment the ROI, they identify an initial region and determine the homogeneity between neighbouring pixels.…”

Section: Introductionmentioning

confidence: 99%

“…This task can be accomplished using a classification algorithm. For example, rule-based methods [23], k-nearest neighbour (KNN) [20] and support vector machine (SVM) [6][7][8][9][10]12,13], are some popular algorithms for false reduction. Similarly, the neural network-based classifiers, such as the artificial neural network (ANN) [11,21] and convolutional neural network (CNN) [24,25], are also found in this step.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Automatic Detection of Pulmonary Nodules using Three-dimensional Chain Coding and Optimized Random Forest

et al. 2020

View full text Add to dashboard Cite

The detection of pulmonary nodules on computed tomography scans provides a clue for the early diagnosis of lung cancer. Manual detection mandates a heavy radiological workload as it identifies nodules slice-by-slice. This paper presents a fully automated nodule detection with three significant contributions. First, an automated seeded region growing is designed to segment the lung regions from the tomography scans. Second, a three-dimensional chain code algorithm is implemented to refine the border of the segmented lungs. Lastly, nodules inside the lungs are detected using an optimized random forest classifier. The experiments for our proposed detection are conducted using 888 scans from a public dataset, and achieves a favorable result of 93.11% accuracy, 94.86% sensitivity, and 91.37% specificity, with only 0.0863 false positives per exam.

show abstract

Automatic detection and segmentation of lung nodules in different locations from CT images based on adaptive α‐hull algorithm and DenseNet convolutional network

Zhang

et al. 2021

Int J Imaging Syst Tech

View full text Add to dashboard Cite

Automatic lung nodules detection and segmentation can assist doctors in better diagnosis and treatment for lung cancer. However, precise detection and segmentation are still challenging, because lung nodules can have different contours or locations and may be attached to other tissues, such as neighboring blood vessel and pleural surface. In this study, an automatic detection and segmentation method for lung nodules in different locations has been developed. First, we apply Otsu thresholding to segment lung parenchyma. Next, a morphological opening operation is carried out to remove blood vessels. Then α-hull operation is proposed to correct lung contours and optimal α values can be acquired adaptively. Finally, DenseNet convolutional network is applied to classify true lung nodules from all nodule candidates. We select the intersection area of at least three radiologists' annotations as ground truth and validate our method on 466 nodules including well-circumscribed, juxta-vascular, juxta-pleural, and pleural tail. Our study not only concentrates on false positive reduction but also evaluates segmentation performance. To give a more comprehensive quantitative evaluation of nodule segmentation, we use evaluation metrics including Jaccard index (JI), dice similar coefficient (DSC), Hausdorff distance, under-segmentation rate, over-segmentation rate, sensitivity, specificity, accuracy, and false positive rate. Overall results are 0.6385 ± 0.

show abstract

3D skeletonization feature based computer-aided detection system for pulmonary nodules in CT datasets

Cited by 48 publications

References 21 publications

A Two-Stage Convolutional Neural Networks for Lung Nodule Detection

A Two-Stage Convolutional Neural Networks for Lung Nodule Detection

Automatic Detection of Pulmonary Nodules using Three-dimensional Chain Coding and Optimized Random Forest

Automatic detection and segmentation of lung nodules in different locations from CT images based on adaptive α‐hull algorithm and DenseNet convolutional network

Contact Info

Product

Resources

About