Journal of Applied Sciences1812-56541812-5662Asian Network for Scientific Information10.3923/jas.2014.2163.2169ZhaoganLuZhiguoBao1220141418By the reasons of the bad accuracy of drawbacks of current
image searching algorithms, the difficulties and uncertainties of image blocking
procedures, one image searching algorithm with the partition of multiple scales
and levels is proposed in this study, which is conducted by the classic wavelet
multiple scale and level decomposition algorithm. According to the algorithm,
the interest objects would be found in the image partitions at different levels
with different scales, where the Gabor texture features, main colors and color
invariant moments were used as the metric of image searching results. Based
the numerical evaluation results with classical histogram and invariant moment
approaches, the image searching accuracy is improved from 66 and 68-77% by the
algorithm with feasible implementation.]]>Abubacker, K.S. and L.K. Indumathi,2010Jing, F., M. Li, H.J. Zhang and B. Zhang,2004Shao, H. and X. Wang,2012Lin, H.J., Y.T. Kao, S.H. Yen and C.J. Wang,2004Pratikakis, I., I. Vanhamel, H. Sahli, B. Gatos and S.J. Perantonis,2006Pujari, J.D. and A.S. Nayak,2010Rahman, M.M., S.K. Antani and G.R. Thoma,2010Sahu, M., M. Shrivastava and M.A. Rizvi,2012Vu, K., K.A. Hua and W. Tavanapong,2003Lu, Y., J. Hu and D. Huang,2006Zhang, M.,2012