This thesis proposes two scalable coding methods for error-diffused bi-level images, which are quality scalability and resolution scalability. In addition, a new method to compress multisymbol data using binary arithmetic coding algorithm is also proposed. The proposed scalable coding methods divide images into cells and then encode their information; the number of dots and the position of dots in cells. The quality scalable coding algorithm scans an image twice. During the first pass, it encodes the number of dots in cells. After then, it encodes the position of dots based on the number of dots in the cell and the information of neighbor cells. For fast computation, the algorithm can be changed into encoding or decoding the S-value and the C-value simultaneously. In order to incorporate resolution scalability with compression algorithm, rehalftoning method is proposed. The S-value used in the proposed quality scalable coding method is considered as gray value of a lower resolution image. The image is halftoned again to make a lower resolution bi-level image. In the proposed scalable coding algorithms, the arithmetic coding algorithm is used to code multiple symbols. In the proposed binary arithmetic coding of multiple symbols, a multiple symbol is represented into bits and the binary arithmetic coder encodes each bit by using the previously encoded bits as the context for the current bit. This method removes multiplications and divisions for calculating intervals and probabilities in the multisymbol arithmetic coding.