Abstract: More requirements for image coding have arisen from the rapid development of multimedia and Internet. In this paper, an overview of major methods already used or under investigation for still image coding based on waveform coding, second-generation coding and fractal coding is first given, and the problem of BRE （bit rate error） in image transmission over noisy channels is analyzed. Also discussed are some techniques to alleviate channel error owing to bit-error in information such as DC coefficients and AC coefficients, desynchronization of block and packet loss from the standpoint of the image codec. In addition, the objectives and coding engine in JPEG-2000, an emerging standard for still image coding, are introduced, and the Maxshift method for ROI coding and the image manipulation based on compressed images in JPEG-2000 are also analyzed. Although the compression efficiency is the most important feature in any compression scheme, other functionalities, such as coding based on content, scalability coding and error resilience etc., should be taken into account when choosing a best coding scheme. Hereby, the future development of still image coding is also analyzed and conjectured.