The chaotic mapping enables the logo pixels rearrange randomly in the specified spatial domain

and the mixed bits of the logo were embedded in the image DCT coefficients. To lay the watermarking information in the notable areas of the features

an algorithm of fuzzy entropy estimation was proposed. The embedding strength was characteristic adaptability with the statistic of the local feature coefficients of the original image. The logo could be detected and reconstructed by a specifically Kalman filter without the original data from the watermarked image

and would provide indication of the copyright existence. Experimental results show that the proposed scheme gives good capability and turns out to be a reliable copyright authentication tool for the digital image.