Local binary pattern texture-based classification of solid masses in ultrasound breast images

Monica M. S. Matsumoto; Chandra M. Sehgal; Jayaram K. Udupa

doi:10.1117/12.911653

25 February 2012 Local binary pattern texture-based classification of solid masses in ultrasound breast images

Monica M. S. Matsumoto, Chandra M. Sehgal, Jayaram K. Udupa

Author Affiliations +

Proceedings Volume 8320, Medical Imaging 2012: Ultrasonic Imaging, Tomography, and Therapy; 83201H (2012) https://doi.org/10.1117/12.911653
Event: SPIE Medical Imaging, 2012, San Diego, California, United States

Abstract

Breast cancer is one of the leading causes of cancer mortality among women. Ultrasound examination can be used to assess breast masses, complementarily to mammography. Ultrasound images reveal tissue information in its echoic patterns. Therefore, pattern recognition techniques can facilitate classification of lesions and thereby reduce the number of unnecessary biopsies. Our hypothesis was that image texture features on the boundary of a lesion and its vicinity can be used to classify masses. We have used intensity-independent and rotation-invariant texture features, known as Local Binary Patterns (LBP). The classifier selected was K-nearest neighbors. Our breast ultrasound image database consisted of 100 patient images (50 benign and 50 malignant cases). The determination of whether the mass was benign or malignant was done through biopsy and pathology assessment. The training set consisted of sixty images, randomly chosen from the database of 100 patients. The testing set consisted of forty images to be classified. The results with a multi-fold cross validation of 100 iterations produced a robust evaluation. The highest performance was observed for feature LBP with 24 symmetrically distributed neighbors over a circle of radius 3 (LBP_24,3) with an accuracy rate of 81.0%. We also investigated an approach with a score of malignancy assigned to the images in the test set. This approach provided an ROC curve with Az of 0.803. The analysis of texture features over the boundary of solid masses showed promise for malignancy classification in ultrasound breast images.

Citation Download Citation

Monica M. S. Matsumoto, Chandra M. Sehgal, and Jayaram K. Udupa "Local binary pattern texture-based classification of solid masses in ultrasound breast images", Proc. SPIE 8320, Medical Imaging 2012: Ultrasonic Imaging, Tomography, and Therapy, 83201H (25 February 2012); https://doi.org/10.1117/12.911653

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

;

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE