Using HOG-LBP features and MMP learning to recognize imaging signs of lung lesions

Li Song; Xiabi Liu; Ling Ma; Chunwu Zhou; Xinming Zhao; Yanfeng Zhao

doi:10.1109/CBMS.2012.6266313

Using HOG-LBP features and MMP learning to recognize imaging signs of lung lesions

Li Song
, Xiabi Liu^*
, Ling Ma
, Chunwu Zhou
, Xinming Zhao
, Yanfeng Zhao

^*Corresponding author for this work

School of Computer Science and Technology

Beijing Institute of Technology
Chinese Academy of Medical Sciences

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

33 Citations (Scopus)

Abstract

This paper proposes an approach to recognize Common Imaging Signs of Lesions (CISLs) in lung CT images. We combine the bag-of-visual-words based on the Histograms of Oriented Gradients (HOG) and the Local Binary Pattern (LBP) to represent regions of interest (ROIs) in lung CT images. Then the Max-Min posterior Pseudo-probabilities (MMP) learning method is applied to recognize the category of the imaging sign contained in each ROI. We conducted the 5-fold cross validation experiments on a set of 696 ROIs captured from real lung CT images. The proposed approach achieved the average sensitivity of 91.8%, the average specificity of 98.5% and the average accuracy of 98%. Furthermore, the HOG-LBP features surpassed individual HOG or LBP as well as the hybrid of LBP and intensity histograms, and the MMP behaved better than the Support Vector Machines (SVMs). These experimental results confirm the effectiveness of our approach.

Original language	English
Title of host publication	Proceedings of the 25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012
DOIs	https://doi.org/10.1109/CBMS.2012.6266313
Publication status	Published - 2012
Event	25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012 - Rome, Italy Duration: 20 Jun 2012 → 22 Jun 2012

Publication series

Name	Proceedings - IEEE Symposium on Computer-Based Medical Systems
ISSN (Print)	1063-7125

Conference

Conference	25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012
Country/Territory	Italy
City	Rome
Period	20/06/12 → 22/06/12

Access to Document

10.1109/CBMS.2012.6266313

Cite this

Song, L., Liu, X., Ma, L., Zhou, C., Zhao, X., & Zhao, Y. (2012). Using HOG-LBP features and MMP learning to recognize imaging signs of lung lesions. In Proceedings of the 25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012 Article 6266313 (Proceedings - IEEE Symposium on Computer-Based Medical Systems). https://doi.org/10.1109/CBMS.2012.6266313

@inproceedings{6453e484b29846b3a0c35c31d65d7951,

title = "Using HOG-LBP features and MMP learning to recognize imaging signs of lung lesions",

abstract = "This paper proposes an approach to recognize Common Imaging Signs of Lesions (CISLs) in lung CT images. We combine the bag-of-visual-words based on the Histograms of Oriented Gradients (HOG) and the Local Binary Pattern (LBP) to represent regions of interest (ROIs) in lung CT images. Then the Max-Min posterior Pseudo-probabilities (MMP) learning method is applied to recognize the category of the imaging sign contained in each ROI. We conducted the 5-fold cross validation experiments on a set of 696 ROIs captured from real lung CT images. The proposed approach achieved the average sensitivity of 91.8\%, the average specificity of 98.5\% and the average accuracy of 98\%. Furthermore, the HOG-LBP features surpassed individual HOG or LBP as well as the hybrid of LBP and intensity histograms, and the MMP behaved better than the Support Vector Machines (SVMs). These experimental results confirm the effectiveness of our approach.",

author = "Li Song and Xiabi Liu and Ling Ma and Chunwu Zhou and Xinming Zhao and Yanfeng Zhao",

year = "2012",

doi = "10.1109/CBMS.2012.6266313",

language = "English",

isbn = "9781467320511",

series = "Proceedings - IEEE Symposium on Computer-Based Medical Systems",

booktitle = "Proceedings of the 25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012",

note = "25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012 ; Conference date: 20-06-2012 Through 22-06-2012",

}

Song, L, Liu, X, Ma, L, Zhou, C, Zhao, X & Zhao, Y 2012, Using HOG-LBP features and MMP learning to recognize imaging signs of lung lesions. in Proceedings of the 25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012., 6266313, Proceedings - IEEE Symposium on Computer-Based Medical Systems, 25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012, Rome, Italy, 20/06/12. https://doi.org/10.1109/CBMS.2012.6266313

Using HOG-LBP features and MMP learning to recognize imaging signs of lung lesions. / Song, Li; Liu, Xiabi; Ma, Ling et al.
Proceedings of the 25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012. 2012. 6266313 (Proceedings - IEEE Symposium on Computer-Based Medical Systems).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Using HOG-LBP features and MMP learning to recognize imaging signs of lung lesions

AU - Song, Li

AU - Liu, Xiabi

AU - Ma, Ling

AU - Zhou, Chunwu

AU - Zhao, Xinming

AU - Zhao, Yanfeng

PY - 2012

Y1 - 2012

N2 - This paper proposes an approach to recognize Common Imaging Signs of Lesions (CISLs) in lung CT images. We combine the bag-of-visual-words based on the Histograms of Oriented Gradients (HOG) and the Local Binary Pattern (LBP) to represent regions of interest (ROIs) in lung CT images. Then the Max-Min posterior Pseudo-probabilities (MMP) learning method is applied to recognize the category of the imaging sign contained in each ROI. We conducted the 5-fold cross validation experiments on a set of 696 ROIs captured from real lung CT images. The proposed approach achieved the average sensitivity of 91.8%, the average specificity of 98.5% and the average accuracy of 98%. Furthermore, the HOG-LBP features surpassed individual HOG or LBP as well as the hybrid of LBP and intensity histograms, and the MMP behaved better than the Support Vector Machines (SVMs). These experimental results confirm the effectiveness of our approach.

AB - This paper proposes an approach to recognize Common Imaging Signs of Lesions (CISLs) in lung CT images. We combine the bag-of-visual-words based on the Histograms of Oriented Gradients (HOG) and the Local Binary Pattern (LBP) to represent regions of interest (ROIs) in lung CT images. Then the Max-Min posterior Pseudo-probabilities (MMP) learning method is applied to recognize the category of the imaging sign contained in each ROI. We conducted the 5-fold cross validation experiments on a set of 696 ROIs captured from real lung CT images. The proposed approach achieved the average sensitivity of 91.8%, the average specificity of 98.5% and the average accuracy of 98%. Furthermore, the HOG-LBP features surpassed individual HOG or LBP as well as the hybrid of LBP and intensity histograms, and the MMP behaved better than the Support Vector Machines (SVMs). These experimental results confirm the effectiveness of our approach.

UR - https://www.scopus.com/pages/publications/84867324825

U2 - 10.1109/CBMS.2012.6266313

DO - 10.1109/CBMS.2012.6266313

M3 - Conference contribution

AN - SCOPUS:84867324825

SN - 9781467320511

T3 - Proceedings - IEEE Symposium on Computer-Based Medical Systems

BT - Proceedings of the 25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012

T2 - 25th IEEE International Symposium on Computer-Based Medical Systems, CBMS 2012

Y2 - 20 June 2012 through 22 June 2012

ER -

Using HOG-LBP features and MMP learning to recognize imaging signs of lung lesions

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this