The novel directional selection based on complex discrete wavelet transform

This paper proposes the novel directional selection based on the complex discrete wavelet transform (CDWT). CDWT is known as a very useful image processing method because of shift-invariant property. Another property of the CDWT is the directional selection that can detect the edges with different direction in images and the directional selection is expected as a geometric feature extraction method. However, the directional selection of the CDWT does not offer various directional features. Therefore, we propose the novel directional selection based on the CDWT and the directional filters We design the directional filters that can offers many directional features references from the curvelet transform and the contourlet Transform.