Effective Pneumothorax Detection for Chest X-Ray Images Using Local Binary Pattern and Support Vector Machine

Yuan-Hao Chan, Yong-Zhi Zeng, Hsien-Chu Wu, Ming-Chi Wu, Hung-Min Sun, Yuan-Hao Chan, Yong-Zhi Zeng, Hsien-Chu Wu, Ming-Chi Wu, Hung-Min Sun

Abstract

Automatic image segmentation and feature analysis can assist doctors in the treatment and diagnosis of diseases more accurately. Automatic medical image segmentation is difficult due to the varying image quality among equipment. In this paper, the automatic method employed image multiscale intensity texture analysis and segmentation to solve this problem. In this paper, firstly, SVM is applied to identify common pneumothorax. Features are extracted from lung images with the LBP (local binary pattern). Then, classification of pneumothorax is determined by SVM. Secondly, the proposed automatic pneumothorax detection method is based on multiscale intensity texture segmentation by removing the background and noises in chest images for segmenting abnormal lung regions. The segmentation of abnormal regions is used for texture transformed from computing multiple overlapping blocks. The rib boundaries are identified with Sobel edge detection. Finally, in obtaining a complete disease region, the rib boundary is filled up and located between the abnormal regions.

Figures

Figure 1
Figure 1
Chest radiographs of (a) normal and (b) pneumothorax.
Figure 2
Figure 2
The flowchart of the proposed SVM-based lung classification.
Figure 3
Figure 3
The results of an example (a) original image, (b) the enhanced image, (c) enhanced by hole-filling image, and (d) image lung region identification.
Figure 4
Figure 4
An example of LBP. This result values encoded as 11001100 by starting from the upper left and reading clockwise.
Figure 5
Figure 5
The ROT function, as the same type in the case of clockwise rotation.
Figure 6
Figure 6
The flowchart of the proposed multiscale intensity texture segmentation.
Figure 7
Figure 7
An example of the pattern generation: (a) the original pixel values, (b) difference value of the neighboring point and center point, (c) result of the Win function, and (d) the pattern value of the LBP function.
Figure 8
Figure 8
The segmentation results: (a) target region, (b) H1 intensity distribution pattern set, (c) H2 intensity distribution pattern set, (d) V1 intensity distribution pattern set, and (e) V2 intensity distribution pattern set.
Figure 9
Figure 9
The results of an example of (a) smooth regions, (b) complex regions, (c) rib boundary regions, (d) intersection region between smooth and complex regions, (e)–(f) final segmentation region, and (g)–(h) original image.
Figure 10
Figure 10
Comparison between the area that is manually depicted of radiologists and the proposed method. (a–c) Segmentation results of the left lung; (d–f) segmentation results of the right lung.
Figure 11
Figure 11
Accuracy, precision, and sensitivity of all dataset image using the proposed method.

References

    1. Noppen M., De Keukeleire T. Pneumothorax. Respiration. 2008;76(2):121–127. doi: 10.1159/000135932.
    1. MacDuff A., Arnold A., Harvey J., on behalf of the BTS Pleural Disease Guideline Group Management of spontaneous pneumothorax: British Thoracic Society pleural disease guideline 2010. Thorax. 2010;65(Supplement 2):ii18–ii31. doi: 10.1136/thx.2010.136986.
    1. Wakai A. P. Spontaneous Pneumothorax. BMJ Clinical Evidence. 2011
    1. Bellaviti N., Bini F., Pennacchi L., et al. Increased incidence of spontaneous pneumothorax in very young people: observations and treatment. Chest. 2016;150(4):p. 560A. doi: 10.1016/j.chest.2016.08.649.
    1. Weber T., Bayer F., Haas W., et al. Investigation of the Signature of Lung Tissue in X-Ray Grating-Based Phase-Contrast Imaging. 2012. .
    1. Cardinale L., Volpicelli G., Lamorte A., Martino J. Revisiting signs, strengths and weaknesses of standard chest radiography in patients of acute dyspnea in the emergency department. Journal of Thoracic Disease. 2012;4(4):398–407. doi: 10.3978/j.issn.2072-1439.2012.05.05.
    1. Doi K., Sanada S. Method and System for Automatic Detection of Ribs and Pneumothorax in Digital Chest Radiographs. US Patent 5,668,888, 1997.
    1. Doi K. Computer-aided diagnosis in medical imaging: historical review, current status and future potential. Computerized Medical Imaging and Graphics. 2007;31(4-5):198–211. doi: 10.1016/j.compmedimag.2007.02.002.
    1. Avni U., Greenspan H., Konen E., Sharon M., Goldberger J. X-ray categorization and retrieval on the organ and pathology level, using patch-based visual words. IEEE Transactions on Medical Imaging. 2011;30(3):733–746. doi: 10.1109/TMI.2010.2095026.
    1. Geva O., Zimmerman-Moreno G., Lieberman S., Konen E., Greenspan H. Pneumothorax detection in chest radiographs using local and global texture signatures. Medical Imaging 2015: Computer-Aided Diagnosis; March 2015; Renaissance Orlando, Florida, at SeaWorld.
    1. Huang S. C., Cheng F. C., Chiu Y. S. Efficient contrast enhancement using adaptive gamma correction with weighting distribution. IEEE Transactions on Image Processing. 2013;22(3):1032–1041. doi: 10.1109/TIP.2012.2226047.
    1. Bush A., Gray H., Denison D. M. Diagnosis of pulmonary hypertension from radiographic estimates of pulmonary arterial size. Thorax. 1988;43(2):127–131. doi: 10.1136/thx.43.2.127.
    1. Otsu N. A threshold selection method from gray-level histograms. IEEE Transactions on Systems, Man, and Cybernetics. 1979;9(1):62–66. doi: 10.1109/TSMC.1979.4310076.
    1. Ahonen T., Hadid A., Pietikainen M. Face description with local binary patterns: application to face recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2006;28(12):2037–2041. doi: 10.1109/TPAMI.2006.244.
    1. Kimpe T., Tuytschaever T. Increasing the number of gray shades in medical display systems—how much is enough? Journal of Digital Imaging. 2007;20(4):422–432. doi: 10.1007/s10278-006-1052-3.
    1. Bouman C., Sauer K. A generalized Gaussian image model for edge-preserving MAP estimation. IEEE Transactions on Image Processing. 1993;2(3):296–310. doi: 10.1109/83.236536.

Source: PubMed

Sponsors and Collaborators

Medical Conditions

Drug Interventions

3
Subscribe