Transcoding proxies are used as intermediaries between generic World Wide W
eb servers and a variety of client devices in order to adapt to the greatly
varying bandwidths of different client communication links and to handle t
he heterogeneity of possibly small-screened client devices. Such transcodin
g proxies can adaptively adjust the amount by which a data stream is reduce
d, using an aggressive lossy compression method (e.g., an image becomes les
s clear, text is summarized) We present an analytical framework for determi
ning whether to transcode and how much to transcode an image for the two ca
ses of store-and-forward transcoding as well as streamed transcoding. These
methods require prediction of transcoding delay, prediction of transcoded
image size (in bytes), and estimation of network bandwidth. We discuss meth
ods of adaptation based on fixed quality as well as fixed delay (automated/
dynamic transcoding). We conclude with a description of the practical adapt
ation policies that have been implemented in our adaptive image transcoding
proxy.