Abstract
A correct mammographic evaluation and interpretation demands a high level of expertise from the observing radiologist, and depends directly on an adequate visual analysis of the findings and correlation of the radiological features extracted from different mammographic views. In this article we present an automatic classification scheme of nodules contained in Regions of Interest (RoIs), extracted from two different mammographic projections (Mid Lateral Oblique and CreaneoCaudal) obtained from the same mammary gland, by means of an ipsilateral information fusion strategy. Once the specialist radiologist selects a Region of Interest in the two mammographic projections, these are characterized by multi-resolution and multi-scale decompositions, for which each RoI is projected on two different spaces defined by the Zernike moments and the Curvelet transform, respectively. Thus, this heterogeneous information is optimally fused by means of a Multiple Kernel Learning strategy built by a Support Vector Machine (SVM) training. The performance of the proposed strategy for classifying malignant and benign nodules was evaluated with respect to a classification scheme based on the analysis of the RoI corresponding to a single projection, using a set of 980 RoI extracted from 490 case studies from the mammography database (DDSM) and 216 RoI extracted from 108 case studies from the INBreast database. The results obtained report a sensitivity of 98.3% and specificity of 94.5% versus a sensitivity of 96.2% and specificity of 93.1% obtained when using multi-resolution features in a single projection. These results suggest that the proposed strategy may be useful in clinical scenarios and may contribute to the training of new radiologists as a second reader.
Translated title of the contribution | Automatic Classification of Mammographic Nodules Based on Fusion of Multi-view Information |
---|---|
Original language | Spanish (Ecuador) |
Title of host publication | Aplicaciones e innovación de la ingeniería en ciencia y tecnología |
Publisher | Editorial Universitaria Abya-Yala |
Pages | 233-264 |
Number of pages | 32 |
ISBN (Print) | 978-9978-10-361-6 |
State | Published - 26 Jun 2019 |
Keywords
- Computer aided diagnosis
- Curvelet transform
- Multiple kernel learning
- Zernike moments
CACES Knowledge Areas
- 8217A Mechatronics