Refinement of lung nodule candidates based on local geometric shape analysis and Laplacian of Gaussian kernels

Saien, Soudeh; Pilevar, Abdol Hamid; Moghaddam, Hamid Abrishami

doi:10.1016/j.compbiomed.2014.09.010

Cited by 34 publications

(11 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some works [ 8 , 9 , 10 , 11 ] used different algorithms for lung reconstruction, among those options some use filling holes algorithm, morphology to define the lungs perimeter or connectivity to reconstruct the lungs volume and verify it as a solid one. Also there are other studies [ 10 , 12 ] with the main task in the threshold decision, the option of this methodology is an adaptive or a minimum error threshold which decides a parameter for lung extraction. The combination of both methods to develop a robust lung extraction represents a significant step for our algorithm of lung nodule detection.…”

Section: Introductionmentioning

confidence: 99%

“…A vast amount of 3D algorithms were proposed targeting spherical volumes mostly because of the similarity to nodules [ 8 , 10 ]. There were also other algorithms focusing in the width of the structures in order to differentiate between tubes and round objects [ 9 , 11 , 12 , 13 , 14 ]. 3D methods require greater processing time and find a noticeable amount of FPs per scan leading towards more demanding classification characteristics.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Auto Diagnostics of Lung Nodules Using Minimal Characteristics Extraction Technique

2016

View full text Add to dashboard Cite

Computer-aided detection (CAD) systems provide useful tools and an advantageous process to physicians aiming to detect lung nodules. This paper develops a method composed of four processes for lung nodule detection. The first step employs image acquisition and pre-processing techniques to isolate the lungs from the rest of the body. The second stage involves the segmentation process using a 2D algorithm to affect every layer of a scan eliminating non-informative structures inside the lungs, and a 3D blob algorithm associated with a connectivity algorithm to select possible nodule shape candidates. The combinations of these algorithms efficiently eliminate the high rates of false positives. The third process extracts eight minimal representative characteristics of the possible candidates. The final step utilizes a support vector machine for classifying the possible candidates into nodules and non-nodules depending on their features. As the objective is to find nodules bigger than 4mm, the proposed approach demonstrated quite encouraging results. Among 65 computer tomography (CT) scans, 94.23% of sensitivity and 84.75% in specificity were obtained. The accuracy of these two results was 89.19% taking into consideration that 45 scans were used for testing and 20 for training. The rate of false positives was 0.2 per scan.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Auto Diagnostics of Lung Nodules Using Minimal Characteristics Extraction Technique

2016

View full text Add to dashboard Cite

show abstract

“…In lung cancer CAD systems, lung nodule detection methods can be categorized into three main categories [ 5 ]: template-based [ 6 – 8 ], segmentation-based [ 9 – 11 ], and classification-based [ 12 – 15 ]. Among the reported existing work, the systems that included a classification component in their structure have performed better than their counterparts.…”

Section: Introductionmentioning

confidence: 99%

Computer-Aided Lung Nodule Recognition by SVM Classifier Based on Combination of Random Undersampling and SMOTE

Yuan

Wei

Zhao

2015

Computational and Mathematical Methods in Medicine

View full text Add to dashboard Cite

In lung cancer computer-aided detection/diagnosis (CAD) systems, classification of regions of interest (ROI) is often used to detect/diagnose lung nodule accurately. However, problems of unbalanced datasets often have detrimental effects on the performance of classification. In this paper, both minority and majority classes are resampled to increase the generalization ability. We propose a novel SVM classifier combined with random undersampling (RU) and SMOTE for lung nodule recognition. The combinations of the two resampling methods not only achieve a balanced training samples but also remove noise and duplicate information in the training sample and retain useful information to improve the effective data utilization, hence improving performance of SVM algorithm for pulmonary nodules classification under the unbalanced data. Eight features including 2D and 3D features are extracted for training and classification. Experimental results show that for different sizes of training datasets our RU-SMOTE-SVM classifier gets the highest classification accuracy among the four kinds of classifiers, and the average classification accuracy is more than 92.94%.

show abstract

“…In the above methods, the multilevel threshold based approaches [2][3][4][5][16][17][18][19][20] usually get better performances than active contour algorithm [6] and supervised voxel classification methods [7]. Equally, it has an obvious disadvantage for segmenting the fuzzy lesion which tends to be the probable modality of malignant nodule such as GGO (Ground Glass Opacity).…”

Section: Introductionmentioning

confidence: 99%

“…Many methods for the lung parenchyma and pulmonary nodules detection have been studied, such as threshold rule based conversion methods [16][17][18][19][20], morphology algorithm [21], genetic algorithm, model based methods [22][23][24], classifier based methods [25] and the enhanced filtering methods [26][27][28]. Li et al use an improved CV Level-Set algorithm for segmenting lung nodules based on SVM (support vector machine) classifier [25].…”

Section: Introductionmentioning

confidence: 99%

A novel lung cancer detection algorithm for CADs based on SSP and Level Set

Zhu

Chun-Hyok

Song

et al. 2017

THC

View full text Add to dashboard Cite

Abstract. The fuzzy degree of lung nodule boundary is the most important cue to judge the lung cancer in CT images. Based on this feature, the paper proposes a novel lung cancer detection method for CT images based on the super-pixels and the level set segmentation methods. In the proposed methods, the super-pixels method is used to segment the lung region and the suspected lung cancer lesion region in the CT image. The super-pixels method and a level set method are used to segment the suspected lung cancer lesion region simultaneously. Finally, the cancer is determined by the difference between results of the two segmentation methods. Experimental results show that the proposed algorithm has a high accuracy for lung cancer detection in CT images. For gross glass nodule, pleural nodule, the vascular nodules and solitary nodules, the sensitivity of the detection algorithm are respectively 91.3%, 96.3%, 80.9% and 82.3%.

show abstract

Refinement of lung nodule candidates based on local geometric shape analysis and Laplacian of Gaussian kernels

Cited by 34 publications

References 43 publications

Auto Diagnostics of Lung Nodules Using Minimal Characteristics Extraction Technique

Auto Diagnostics of Lung Nodules Using Minimal Characteristics Extraction Technique

Computer-Aided Lung Nodule Recognition by SVM Classifier Based on Combination of Random Undersampling and SMOTE

A novel lung cancer detection algorithm for CADs based on SSP and Level Set

Contact Info

Product

Resources

About