Published March 24, 2007 | Version 1196

Analysis of Sonographic Images of Breast

Description

Ultrasound images are very useful diagnostic tool to distinguish benignant from malignant masses of the breast. However, there is a considerable overlap between benignancy and malignancy in ultrasonic images which makes it difficult to interpret. In this paper, a new noise removal algorithm was used to improve the images and classification process. The masses are classified by wavelet transform's coefficients, morphological and textural features as a novel feature set for this goal. The Bayesian estimation theory is used to classify the tissues in three classes according to their features.

Files

1196.pdf

Files (548.0 kB)

Name Size
md5:113bc30ea1d4fd4e4071d4af50fc7303
548.0 kB Preview Download