Key points are not available for this paper at this time.
Current computer aided diagnosis (CADx) software for digital mammography relies mainly on 2D techniques. With the emergence of three-dimensional (3D) breast imaging modalities such as breast Computed Tomography (BCT), there is an opportunity to analyze 3D features in the classification of calcifications. We previously reported our initial work on automated 3D feature detection and classification based on morphological descriptions for single microcalcifications within clusters 1. In this work, we propose the expansion of the 3D classification methods to include novel microcalcification morphological feature detection such as including more morphological classes and replacing the 2D Radon transform by a 3D Radon transform. Results show that the classification rate improved compared to the previously reported results from a total of 546 to 559 consistently classified calcifications out of 635 total calcifications. This slight improvement is due to the use of the 3D Radon transform and incorporating methods to detect two classes not previously implemented. Future work will focus on adding feature detection and classification of cluster patterns.
Alquran et al. (Fri,) studied this question.