Article: Automatic detection of the tumour on mammogram images based on hidden Markov and active contour with quasi-automatic initialisation Journal: International Journal of Medical Engineering and Informatics (IJMEI) 2017 Vol.9 No.4 pp.316 - 331 Abstract: The area of tumour's detection and removal is a very active research within the field of medical imaging. In the present work, we present an automatic method for tumour's detection in mammography images. The proposed approach is to construct a detection pattern which starts with the Otsu's method: the thresholding step, followed by estimating the number of classes based on the local binary pattern (LBP) technique. To automate the initialisation task, we proposed to apply the classification by the k-means dynamic improved by Markov's method. The tumour's image is the result of the maximum correlation. A second contribution which is based on active contours gradient vector flow (GVF) with quasi-automatic initialisation applied on the structure that resulted from the structure/texture decomposition of the image to classify. The experimental results show the quality and automation of tumour's detection in medical images in comparison to literature methods. Inderscience Publishers - linking academia, business and industry through research

Title: Automatic detection of the tumour on mammogram images based on hidden Markov and active contour with quasi-automatic initialisation

Authors: Soukaina El Idrissi El Kaitouni; Abdelghafour Abbad; Hamid Tairi

Addresses: LIIAN, Department of Informatics Faculty of Science Dhar-Mahraz, University of Sidi Mohamed Ben Abdellah, P.B 1796 Atlas-Fez, Morocco ' LIIAN, Department of Informatics Faculty of Science Dhar-Mahraz, University of Sidi Mohamed Ben Abdellah, P.B 1796 Atlas-Fez, Morocco ' LIIAN, Department of Informatics Faculty of Science Dhar-Mahraz, University of Sidi Mohamed Ben Abdellah, P.B 1796 Atlas-Fez, Morocco

Abstract: The area of tumour's detection and removal is a very active research within the field of medical imaging. In the present work, we present an automatic method for tumour's detection in mammography images. The proposed approach is to construct a detection pattern which starts with the Otsu's method: the thresholding step, followed by estimating the number of classes based on the local binary pattern (LBP) technique. To automate the initialisation task, we proposed to apply the classification by the k-means dynamic improved by Markov's method. The tumour's image is the result of the maximum correlation. A second contribution which is based on active contours gradient vector flow (GVF) with quasi-automatic initialisation applied on the structure that resulted from the structure/texture decomposition of the image to classify. The experimental results show the quality and automation of tumour's detection in medical images in comparison to literature methods.

Keywords: detection; tumours; mammogram image; Otsu thresholding; local binary pattern; LBP; k-means; Markov; structure/texture decomposition; GVF active contours.

DOI: 10.1504/IJMEI.2017.086898

International Journal of Medical Engineering and Informatics, 2017 Vol.9 No.4, pp.316 - 331

Accepted: 25 Nov 2016
Published online: 02 Oct 2017 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Automatic detection of the tumour on mammogram images based on hidden Markov and active contour with quasi-automatic initialisation

Keep up-to-date