Quantized Local Edge Distribution: A Descriptor for B-mode Ultrasound Images
- 1.9k Downloads
This paper presents an illumination invariant, histogram equalization invariant, rotation-robust and spatially stable texture descriptor for B-mode ultrasound images. We design a new edge-encoding descriptor that captures edge distributions of ultrasound textures. The distribution of edges categorized by their strength forms a signature of a specific textural pattern. Oriented edges are first quantized into different levels of salience according to local contrast and then aggregated to polar bins. A distance function that incorporates with our descriptor for effective texture comparison is introduced. The performance of the proposed descriptor is evaluated by various experiments.
KeywordsB-mode ultrasound image image descriptor edge statistic
Unable to display preview. Download preview PDF.
- 1.Shankar, P.M.: A general statistical model for ultrasonic backscattering from tissues. IEEE Trans. on Ultrason., Ferroelect. and Freq. Control 47(3) (2000)Google Scholar
- 2.Bouhlel, N., Sevestre-Ghalila, S.: Nakagami markov random field as texture model for ultrasound rf envelope image. Computers in Biology and Medicine 39(6) (2009)Google Scholar
- 8.Haralick, R.M., Dinstein, Shanmugam, K.: Textural features for image classification. IEEE Trans. on Systems, Man, and Cybernetics (3), 610–621 (1973)Google Scholar
- 9.Dalal, N., Triggs, B.: Histograms of oriented gradients for human detection. In: CVPR, pp. 886–893 (2005)Google Scholar
- 10.Rosin, P.L.: Edges: Saliency measures and automatic thresholding. In: Machine Vision and Application, pp. 139–159 (1997)Google Scholar
- 11.Qian, R., Huang, T.: A two-dimensional edge detection scheme for general visual processing. Pattern Recognition 1, 595–598 (1994)Google Scholar