This paper presents a composite blind digital watermarking technique, CompMark, to hide a visually meaningful grayscale logo in a host image. The multi-resolution fusion principles are used to embed the grayscale logo in perceptually significant blocks in wavelet subbands of the host image. A modulus approach is further used to embed a binary counterpart of the logo in the approximation subband. The extraction process combines both extracted grayscale and binary logos to obtain a better, recognizable logo at the receiver. The watermark detection decision is based on either the correlation between the thresholded extracted and embedded logos or the visual similarity. Experimental results demonstrate our scheme is robust against compression, white noise addition, histogram equalization, and image filtering techniques. In addition, it performs better than a peer blind scheme, XFuseMark.