Segmentation of Lung Lobes in Isotropic CT Images Using Wavelet Transformation

In anatomy, the lung can be divided by lung fissures into several pulmonary lobe units with specific functions. Identifying the lung lobes and the distribution of various diseases among different lung lobes from CT images is important for disease diagnosis and tracking after recovery. In order to solve the problems of low tubular structure segmentation accuracy and long algorithm time in segmenting lung lobes based on lung anatomical structure information, we propose a segmentation algorithm based on lung fissure surface classification using a point cloud region growing approach. We cluster the pulmonary fissures, transformed into point cloud data, according to the differences in the pulmonary fissure surface normal vector and curvature estimated by principal component analysis. Then, a multistage spline surface fitting method is used to fill and expand the lung fissure surface to realize the lung lobe segmentation. The proposed approach was qualitatively and quantitatively evaluated on a public dataset from Lobe and Lung Analysis 2011 (LOLA11), and obtained an overall score of 0.84. Although our approach achieved a slightly lower overall score compared to the deep learning based methods (LobeNet_V2 and V-net), the inter-lobe boundaries from our approach were more accurate for the CT images with visible lung fissures.

Download Full-text

A computer-aided diagnosis system for lung nodule detection in pulmonary CT images based on LS-SVM and wavelet transformation

10.32920/ryerson.14655537.v1 ◽

2021 ◽

Author(s):

Omid Talakoub

Keyword(s):

Wavelet Transformation ◽

Ct Images ◽

Computer Aided Diagnosis ◽

Regions Of Interest ◽

Support Vector ◽

Cad Systems ◽

Cad System ◽

Computer Aided ◽

Spherical Structures ◽

Aided Diagnosis

One of the most important areas of biomedical engineering is medical imaging. Fully automated schemes are currently being explored as Computer-Aided Diagnosis (CAD) systems to provide a second opinion to medical professionals; of these systems, abnormal region detector in medical images is one of the most critical CAD systems in development. The primary motivation in using these systems is due to the fact that reading an enormous number of images is a time-consuming task for the radiologist. This task can be sped up by using a CAD system which highlights abnormal regions of interest. Low false positive rates and high sensitivity are essential requirement[s] of such a system. The initial requirement of processing any organ is an accurate segmentation of the target of interest in the images. A segmentation method based on the wavelet transformation is proposed which accurately extracts lung regions in the thoracic CT images. After this step, an Aritifical Intelligence system, known as Least Squares Support Vector Machine (LS-SVM), is employed to classify nodules within the regions of interest. It is a well known fact that the lung nodules, except the pleural nodules, are mostly spherical structures whereas other structures including blood vessels are shaped as other structures such as tubular. Therfore, an enhancment filter is developed in which spherical structures are accentuated. Processing three different real databases revealed that the proposed system has reached the objective of a CAD system to provide reliable opinion for the doctors in the diagnosis fashion.

Download Full-text

A computer-aided diagnosis system for lung nodule detection in pulmonary CT images based on LS-SVM and wavelet transformation

10.32920/ryerson.14655537 ◽

2021 ◽

Author(s):

Omid Talakoub

Keyword(s):

Wavelet Transformation ◽

Ct Images ◽

Computer Aided Diagnosis ◽

Regions Of Interest ◽

Support Vector ◽

Cad Systems ◽

Cad System ◽

Computer Aided ◽

Spherical Structures ◽

Aided Diagnosis

One of the most important areas of biomedical engineering is medical imaging. Fully automated schemes are currently being explored as Computer-Aided Diagnosis (CAD) systems to provide a second opinion to medical professionals; of these systems, abnormal region detector in medical images is one of the most critical CAD systems in development. The primary motivation in using these systems is due to the fact that reading an enormous number of images is a time-consuming task for the radiologist. This task can be sped up by using a CAD system which highlights abnormal regions of interest. Low false positive rates and high sensitivity are essential requirement[s] of such a system. The initial requirement of processing any organ is an accurate segmentation of the target of interest in the images. A segmentation method based on the wavelet transformation is proposed which accurately extracts lung regions in the thoracic CT images. After this step, an Aritifical Intelligence system, known as Least Squares Support Vector Machine (LS-SVM), is employed to classify nodules within the regions of interest. It is a well known fact that the lung nodules, except the pleural nodules, are mostly spherical structures whereas other structures including blood vessels are shaped as other structures such as tubular. Therfore, an enhancment filter is developed in which spherical structures are accentuated. Processing three different real databases revealed that the proposed system has reached the objective of a CAD system to provide reliable opinion for the doctors in the diagnosis fashion.

Download Full-text

Segmentation of lung lobes in clinical CT images

International Journal of Computer Assisted Radiology and Surgery ◽

10.1007/s11548-008-0153-5 ◽

2008 ◽

Vol 3 (1-2) ◽

pp. 151-163 ◽

Cited By ~ 7

Author(s):

Qiao Wei ◽

Yaoping Hu ◽

John H. MacGregor ◽

Gary Gelfand

Keyword(s):

Ct Images ◽

Lung Lobes

Download Full-text

Automatic recognition of lung lobes and fissures from multislice CT images

10.1117/12.534499 ◽

2004 ◽

Cited By ~ 11

Author(s):

Xiangrong Zhou ◽

Tatsuro Hayashi ◽

Takeshi Hara ◽

Hiroshi Fujita ◽

Ryujiro Yokoyama ◽

...

Keyword(s):

Multislice Ct ◽

Ct Images ◽

Automatic Recognition ◽

Lung Lobes

Download Full-text

High-resolution computed tomography evaluation of the bronchial lumen to vertebral body diameter and pulmonary artery to vertebral body diameter ratios in anesthetized ventilated normal cats

Journal of Feline Medicine and Surgery ◽

10.1177/1098612x16668931 ◽

2016 ◽

Vol 19 (10) ◽

pp. 1007-1012 ◽

Cited By ~ 4

Author(s):

Tekla M Lee-Fowler ◽

Robert C Cole ◽

A Ray Dillon ◽

D Michael Tillson ◽

Rachel Garbarino ◽

...

Keyword(s):

Young Adult ◽

High Resolution ◽

Pulmonary Artery ◽

Vertebral Body ◽

Ct Images ◽

Lung Lobe ◽

Body Diameter ◽

Bronchial Lumen ◽

Pulmonary Artery Diameter ◽

Lung Lobes

Objectives Bronchial lumen to pulmonary artery diameter (BA) ratio has been utilized to investigate pulmonary pathology on high-resolution CT images. Diseases affecting both the bronchi and pulmonary arteries render the BA ratio less useful. The purpose of the study was to establish bronchial lumen diameter to vertebral body diameter (BV) and pulmonary artery diameter to vertebral body diameter (AV) ratios in normal cats. Methods Using high-resolution CT images, 16 sets of measurements (sixth thoracic vertebral body [mid-body], each lobar bronchi and companion pulmonary artery diameter) were acquired from young adult female cats and 41 sets from pubertal female cats. Results Young adult and pubertal cat BV ratios were not statistically different from each other in any lung lobe. Significant differences between individual lung lobe BV ratios were noted on combined age group analysis. Caudal lung lobe AV ratios were significantly different between young adult and pubertal cats. All other lung lobe AV ratios were not significantly different. Caudal lung lobe AV ratios were significantly different from all other lung lobes but not from each other in both the young adult and pubertal cats. Conclusions and relevance BV ratio reference intervals determined for individual lung lobes could be applied to both young adult and pubertal cats. Separate AV ratios for individual lung lobes would be required for young adult and pubertal cats. These ratios should allow more accurate evaluation of cats with concurrent bronchial and pulmonary arterial disease.

Download Full-text

Fully Automatic Detection and Segmentation Approach for Juxta-Pleural Nodules From CT Images

International Journal of Healthcare Information Systems and Informatics ◽

10.4018/ijhisi.20210401.oa5 ◽

2021 ◽

Vol 16 (2) ◽

pp. 87-104

Author(s):

Vijayalaxmi Mekali ◽

Girijamma H. A.

Keyword(s):

Lung Parenchyma ◽

Ct Images ◽

Experimental Result ◽

Computational Time ◽

Detection Rates ◽

Iterative Thresholding ◽

Automated Method ◽

Segmentation Approach ◽

Thresholding Algorithm ◽

Lung Lobes

Early detection of all types of lung nodules with different characters in medical modality images using computer-aided detection is the best acceptable remedy to save the lives of lung cancer sufferers. But accuracy of different types of nodule detection rates is based on chosen segmented procedures for parenchyma and nodules. Separation of pleural from juxta-pleural nodules (JPNs) is difficult as intensity of pleural and attached nodule is similar. This research paper proposes a fully automated method to detect and segment JPNs. In the proposed method, lung parenchyma is segmented using iterative thresholding algorithm. To improve the nodules detection rate separation of connected lung lobes, an algorithm is proposed to separate connected left and right lung lobes. The new method segments JPNs based on lung boundary pixels extraction, concave points extraction, and separation of attached pleural from nodule. Validation of the proposed method was performed on LIDC-CT images. The experimental result confirms that the developed method segments the JPNs with less computational time and high accuracy.

Download Full-text