A Feature-based Tumor Detection for MRI Breast Imaging

In this paper, we propose a feature-based scheme composed of the intra-slice, texture and inter-slice analyses to achieve tumor region identification in MRI breast images. Our intra-slice analysis evaluates the intensity and size information of candidate regions. To find a precise region, a region growing algorithm is proposed based on ellipse fitness. In the texture analysis, some texture features are extracted and combined with a neural network to reduce the false alarms. The inter-slice analysis is based on the continuity characteristic to verify the static behavior of tumor regions. The experimental results show that our proposed scheme can correctly identify tumor regions.