WAVELET BASED SEGMENTATION USING OPTIMAL STATISTICAL FEATURES ON BREAST IMAGES

Abstract

Elastography is the emerging imaging modality that analyzes the stiffness of the tissue for detecting and classifying breast tumors. Computer-aided detection speeds up the diagnostic process of breast cancer improving the survival rate. A multiresolution approach using Discrete wavelet transform is employed on real time images, using the low-low (LL), low-high (LH), high-low (HL), and high-high (HH) sub-bands of Daubechies family. Features are extracted, selected and then finally segmented by K-means clustering algorithm. The proposed work can be extended to Classification of the tumors.

Keywords

• Image Recognition
• Spiking Neuron
• FPGA
• Artificial Neural Networks
• Feature Extraction