MSMT grant 1M6798555601 DAR

 
Diagnostic Enhancement of Screening Mammograms by Means of Local Texture Models
  Jiri Grim     Petr Somol     Michal Haindl     Jan Danes

This page shows examples from the DDSM database of the University of South Florida illustrating the method of preprocessing of screening mammograms as proposed in the paper published in the IEEE Transactions on Image Processing, Vol.18, NO.4, April 2009.

Abstract:
Statistically based preprocessing of screening mammograms is proposed with the aim to in-crease the diagnostic conspicuity of mammographic lesions. We estimate first the local statis-tical texture model of a single mammogram as a joint probability density of grey levels in a suitably chosen search window. The probability density in the form of multivariate Gaussian mixture is estimated from data obtained by pixel-wise scanning the mammogram with the search window. In the second phase we evaluate the estimated density at each position of the window and display the corresponding log-likelihood value as grey level at window center. Light grey levels correspond to the typical parts of the image and the dark values reflect unusual places. The resulting log-likelihood image closely correlates with fine structural details of the original mammogram and facilitates diagnostic interpretation of suspect abnormalities.

Index Terms:
Screening mammography, computer-aided detection, texture information, local statistical model, Gaussian mixture, EM algorithm, log- likelihood image.

Sample Results

B_3017_1
Ground Truth		Original		Processed
Left MLO: Lesion type: mass, lobulated shape, ill-defined margins. Left CC: Lesion type: mass, shape: lobulated-architectural distortion, microlobulated-spiculated margins.
B_3020_1
Ground Truth		Original		Processed
Right MLO: Lesion type: mass, lobulated shape, circumscribed-ill defined margins. Right CC: Lesion type: mass, lobulated shape, circumscribed-ill defined margins.
B_3056_1
Ground Truth		Original		Processed
Left MLO: Lesion type: mass, shape: focal-asymmetric density, margins n/a. Left CC: Lesion type: mass, shape: focal-asymmetric density, margins n/a.
C_0001_1
Ground Truth		Original		Processed
Right MLO: Lesion type: mass, irregular shape, spiculated margins. Right CC: Lesion type: mass, irregular shape, spiculated margins.
C_0143_1
Ground Truth		Original		Processed
Right MLO: Lesion type: mass, irregular shape, ill-defined margins. Right CC: Lesion type: mass, round shape, speculated margins.
D_4163_1
Ground Truth		Original		Processed
Right MLO: Lesion type: calcification, punctate type, clustered distribution. Right CC: Lesion type: calcification, punctate type, clustered distribution.

Further information can be found in the paper:

	Grim, J. - Somol, P. - Haindl, M. - Danes, J. Computer-Aided Evaluation of Screening Mammograms Based on Local Texture Models IEEE Transactions on Image Processing, Vol.18, NO.4, April 2009 pp. 765 - 773.

---

Back to the demonstration page.