This paper presents a new secret image sharing method using multiwavelet transform in a general framework with the associated pre- and post-filtering to achieve an efficient representation of secret images at a low bit rate. It adopts a set partitioning in hierarchical trees (SPIHT) codec to enhance compactness of the data for processing in an image-sharing scheme. Because multiple data channels are involved in the multiwavelet transform, the standard multiresolution representation of the multiwavelet transform does not have the same data structure as that of the scalar wavelet transform for SPIHT encoding. We propose two structure rearrangement schemes, namely, merging and splitting schemes, for adaptation to the SPIHT. Experimental results have demonstrated the effectiveness and high performance of the proposed method on attaining small shadow size and high peak-signal-to-noise ratio versus bit rate.