Texture image retrieval using adaptive directional wavelet transform
Yuichi Tanaka, Madoka Hasegawa, Shigeo Kato · 2009
In this paper, we present an application of adaptive directional wavelet transform (WT) for content-based texture image retrieval. The adaptive directional WT is an alternative of the traditional separable WT for image coding since it is able to transform along diagonal orientations as well as horizontal/vertical orientations and keeps perfect reconstruction property by lifting factorization. We use its transform direction data to increase the image retrieval ratio. The proposed approach obtains higher retrieval ratio than the separable WT and the contourlet transform.