Prof. Brian L. Evans
Collaboration with Mr. Niranjan Damera-Venkata and Dr. Thomas D. Kite
Laboratory for Image and
The University of Texas at Austin, Austin, TX 78712-1084
Digital halftoning is the process by which a continuous-tone image is converted to a binary image, or halftone, for printing or display on binary devices. With the emerging JBIG2 standard, compression of halftone data has received special attention. Of the three halftone compression methods being considered for JBIG2, the method with the highest compression ratio converts the binary halftone to a grayscale image and compresses the grayscale image. Using this approach, the halftone can be compressed up to 16:1 without noticeable loss of quality. The coded image may then be rehalftoned at the decoder using a different halftoning method than was used for encoding. The decoder has the freedom to choose the halftoning method optimized for its printing or display devices.
The emerging JBIG2 standard has two modes: scan and random. The random mode, which is targeted for document retrieval and processing, would be implemented on desktop computers. The scan mode, which is targeted for printers, scanners, and fax machines, would be implemented using embedded hardware/software. Embedded implementations impose limitations on memory and algorithm complexity. The talk describes halftoning, inverse halftoning, and rehalftoning algorithms suitable for embedded implementations of JBIG encoders and decoders.
Talk in PDF format