Center for Embedded Network Sensing

We examine energy tradeoffs in image communication on highly resource-constrained platforms. A JPEG encoder is implemented on Cyclops which allows a compression factor of six for commonly-used 64x64 and 128x128 images without visible quality degradation. Although computation cost is increased due to the encoding step, the communication cost is much reduced due to the reduction in data size. Results indicate significant energy and speed improvements compared to transmitting uncompressed images.