Image feature extraction for mass detection in digital mammography: Influence of wavelet analysis

Rationale and objectives. The objective of this work is to evaluate the imp ortance of image preprocessing, using multiresolution and multiorientation wavelet transforms (WTs), on the performance of a previously reported compu ter assisted diagnostic (CAD) method for breast cancer screening, using dig ital mammography. Method: An analysis of the influence of WTs on image feat ure extraction for mass detection is achieved by comparing the discriminant ability of features extracted with and without the wavelet-based image pre processing using computed ROC. Thr ee indexes are proposed to assess the se gmentation of the mass area with comparison to the ground truth. Data was a nalyzed on the region-of-interest (ROI) database that included mass and nor mal regions from digitized mammograms with the ground truth. Results: The m etrics for the measurement of segmentation of the mass clearly demonstrated the importance of image preprocessing methods. Similarly, the relative imp rovement in performance was observed in feature extraction based on the eva luation of the ROC curves, where the Az values are increased, for example, from 0.71 to 0.75 for a pixel intensity feature and from 0.72 to 0.85 for a morphological feature of the Normalized Deviation of Radial Length. The im provement, therefore, depends on the feature characteristics, being large f or boundary-related features while small for intensity-related features. Co nclusion: The use of image preprocessing modules using wavelet transforms r esults in a significant improvement in feature extraction for the previousl y proposed CAD detection method. We are therefore exploring additional impr ovement in wavelet-based image preprocessing methods, including adaptive me thods, to achieve a further improvement in performance and an evaluation on lar ger image databases. (C) 1999 American Association of Physicists in Me dicine. [S0094-2405(99)01603-X].