Orthogonal and bi-orthogonal wavelets can represent an image in only one way, whereas redundant systems can represent an image in a multitude of ways. We discuss how one can utilize this for image compression using wavelet frames, the advantages of doing so as well as the difficulties that arise.