An algorithm for segmentation of lung ROI by mean-shift clustering combined with multi-scale HESSIAN matrix dot filtering

Abstract

A new algorithm for segmentation of suspected lung ROI (regions of interest) by mean-shift clustering and multi-scale HESSIAN matrix dot filtering was proposed. Original image was firstly filtered by multi-scale HESSIAN matrix dot filters, round suspected nodular lesions in the image were enhanced, and linear shape regions of the trachea and vascular were suppressed. Then, three types of information, such as, shape filtering value of HESSIAN matrix, gray value, and spatial location, were introduced to feature space. The kernel function of mean-shift clustering was divided into product form of three kinds of kernel functions corresponding to the three feature information. Finally, bandwidths were calculated adaptively to determine the bandwidth of each suspected area, and they were used in mean-shift clustering segmentation. Experimental results show that by the introduction of HESSIAN matrix of dot filtering information to mean-shift clustering, nodular regions can be segmented from blood vessels, trachea, or cross regions connected to the nodule, non-nodular areas can be removed from ROIs properly, and ground glass object (GGO) nodular areas can also be segmented. For the experimental data set of 127 different forms of nodules, the average accuracy of the proposed algorithm is more than 90%.

Keywords

HESSIAN matrix / multi-scale dot filtering / mean-shift clustering / segmentation of suspected areas / lung computer-aided detection/diagnosis

References

[1] ChengY. Z.. Mean Shift, mode seeking, and clustering [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1995, 17(8): 790-799

[2] ComaniciuD., MeerP.. Mean shift: A robust approach toward feature space analysis [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2002, 24(5): 603-619

[3] OzertemU., ErdogmusD., JenssenR.. Mean shift spectral clustering [J]. Pattern Recognition, 2008, 41(6): 1924-1938

[4] RaoS., de Medeiros MartinsW., PríncipeC.. Mean shift: An information theoretic perspective [J]. Pattern Recognition Letters, 2009, 30: 222-230

[5] Chang-ChienS. J., HungW.-L., YangM.-S.. On mean shift-based clustering for circular data [J]. Soft Computing-A Fusion of Foundations, Methodologies and Applications, 2012, 16(6): 1043-1060

[6] MingD., CiT., CaiH., LiL., QiaoC., DuJ.. Semivariogram-based spatial bandwidth selection for remote sensing image segmentation with mean-shift algorithm [J]. IEEE Geoscience and Remote Sensing Letters, 2012, 9(5): 813-817

[7] NiemeijerM., LoogM., AbràmoffM. D., ViergeverM. A., ProkopM., van GinnekenB.. On combining computer-aided detection systems [J]. IEEE Transaction on Medical Imaging, 2011, 30(2): 215-223

[8] GigerM. L., ChanH.-P., BooneJ.. History and status of CAD and quantitative image analysis: The role of medical physics and AAPM [J]. Medical Physics, 2008, 35(12): 5799-5820

[9] BellottiR., CarloF. D., GarganoG., TangaroS., CascioS., CatanzaritiD., CerelloE., CheranP., DeloguS. C., De-MitriP., FucheriI., GrossoC., ReticoD., SquarciaA., TommasiS., GolosioE.. A CAD system for nodule detection in low-dose lung CTs based on region growing and a new active contour model [J]. Medical Physics, 2007, 34(12): 4901-4910

[10] MessayT., HardieR. C., RogersS. K.. A new computationally efficient cad system for pulmonary nodule detection in CT imagery [J]. Medical Image Analysis, 2010, 14(3): 390-406

[11] ZhaoB. S., GamsuG., GinsburgM. S., JiangL., SchwarzL. H.. Automatic detection of small lung nodules on CT utilizing a local density maximum algorithm [J]. Journal of Applied Clinical Medical Physics, 2003, 4(3): 248-260

[12] KoJ. P., NaidichD. P.. Computer-aided diagnosis and the evaluation of lung disease [J]. Journal of Thoracic Imaging, 2004, 19: 136-155

[13] YeX., LinX.. Jamshid Dehmeshki. Shape-based computer-aided detection of lung nodules in thoracic CT Images [J]. IEEE Transactions on Biomedical Engineering, 2009, 5(7): 1810-1820

[14] PaikD. S., BeaulieuC. F., RubinG. D., AcgrB., JeffeyrbJ. R., YeeJ., DeY. J.. Surface normal overlap: A computer-aided detection algorithm with application to colonic polyps and lung nodules in helical CT [J]. IEEE Transaction on Medical Imaging, 2004, 23(6): 661-675

[15] KimH. S., MaekadoM., TanJ. K., IshikawaS.. Ground-glass opacity detection by using correlation between successive slice images [J]. International Journal on Artificial Intelligence Tools, 2007, 16(4): 583-592

[16] DiciottiS., LombardoS., FalchiniM., PicozziG., MascalchiM.. Automated segmentation refinement of small lung nodules in ct scans by local shape analysis [J]. IEEE Transactions On Biomedical Engineering, 2007, 58(12): 3418-3428

[17] AntonelliM., CococcioniM., LazzeriniB., MarcelloniF.. Computer-aided detection of lung nodules based on decision fusion techniques [J]. Pattern Analysis and Application, 2011, 14(3): 295-310

[18] Suárez-CuencaJ., TahocesP., SoutoM., LadoM., Remy-JardinM., RemyJ.. Application of the iris filter for automatic detection of pulmonary nodules on computed tomography images [J]. Computers in Biology and Medicine, 2009, 39(10): 921-933

[19] DehmeshkiJ., YeX., LinX.. Automated detection of lung nodules in CT images using shape-based genetic algorithm [J]. Computerized Medical Imaging and Graphics, 2007, 31(6): 408-417

[20] LiQ., SoneS., DoiK.. Selective enhancement filters for nodules, vessels, and airway walls in two- and three-dimensional CT scans [J]. Medical Physics, 2003, 30(8): 2040-2051

[21] LiR.Research on segmentation algorithms of regions of interest for pulmonary lesions [D], 2009ShenyangNortheastern University

[22] WeiY., YuQ., JiaT.. Detection algorithm for suspected lesion in CT images of lung cancer CAD [J]. Chinese Journal of Scientific Instrument, 2009, 30(1): 1-6

[23] ArmatoS. G., MclennanG., Mcnitt-GrayM. F.. Lung image database consortium: developing a resource for the medical imaging research community [J]. Radiology, 2004, 232(3): 739-748