Color Correcting JPEG Compressed Images

Abstract

Many high end printing systems store and/or transmit compressed images to save bandwidth, memory, and disk space. Prior to printing, the system will decompress and then color correct the image to compensate for the characteristics of a particular printer. In the conventional architecture, the decompression step converts the image from some encoded format (e.g. JPEG with Huffman encoding) into the spatial domain. In this domain, the color image may be stored in some standard device independent color space such as IJG/CCIR YCrCb, SMPTE RGB or Kodak Photo Ycc. The ensuing color correction transforms colors from this space to device coordinates. Our proposed technique uses information about the JPEG compressed format to gain efficiency in the color transformation.