The gop parallel image processor

  • Goesta H. Granlund
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 109)


Images contain a great deal of information which requires large processing capabilities. For that purpose fast image processors have been developed. So far they have mainly dealt with processing of binary images obtained by thresholding gray scale images. For segmentation of images having more subtle features such as noisy lines or edges, texture, color, etc. more elaborate procedures have to be used.

A new type of image processor, GOP (General Operator Processor), has been developed. It can work on gray scale or color images of any size, where it uses a combination of local and global processing which makes it possible to detect faint lines or edges. It also produces texture descriptions which can be integrated in the processing to enable segmentation based upon textural features. The processor can be used for classification and segmentation using simultaneously up to 16 different transforms or representations of an image. Feedback controlled processing and relaxation operations can also be implemented with the processor.

The GOP processor can be connected to any system for picture processing where it speeds up the processing by a factor of 200–1000, dependent upon the situation. Processing of a 512×512 image with a 3×3 operator takes approximately 0,5 seconds in the processor.


  1. 1.
    Kidode M, Asada H, Watanabe S: Local Parallel Pattern Processor, PPP Toshiba Review No. 125, 1980.Google Scholar
  2. 2.
    Lougheed R.M, McCubbrey D.L, Sternberg, S.R: Cytocomputers: Architectures for Parallel Image Processing. Environmental Research Institute of Michigan Ann Arbor, Michigan, 1980.Google Scholar
  3. 3.
    Kruse B: System Architecture for Image Analysis Structured Computer Vision Eds S. Tanimoto and A. Klinger Academic Press 1980, pp 169–212.Google Scholar
  4. 4.
    Gemmar P, Ischen H, Luetjen K: FLIP: A Multiprocessor System for Image Processing. Report from Forschungsinstitut fuer Informationsverarbeitung und Mustererkennung. Karlsruhe, Germany, 1980.Google Scholar
  5. 5.
    Gerritsen F.A, Aardema L.G: DIP-1: A Fast, Flexible and Dynamically Microprogrammable Pipelined Image Processor. Report from Delft University of Technology, Delft, Netherlands, 1980.Google Scholar
  6. 6.
    Briggs F.A, Fu K.S, Hwang K, and Patel J.H: PM4 — A Reconfigurable Multiprocessor System for Pattern Recognition and Image Processing. AFIPS-Conference Proceedings, Vol. 48, 1979.Google Scholar
  7. 7.
    Van Daele I, De Roo J, Vanderheydt L, Oosterlinck A, Van den Berghe H: Image Computer Configuration with Video Rate Processing Capabilities. Proc. First Scandinavian Conference on Image Analysis, Linköping, Sweden, 1980.Google Scholar
  8. 8.
    Reeves A.P: A Systematically Designed Binary Array Processor. IEEE Trans. Comp, Vol. C-29, No. 4, 1980.Google Scholar
  9. 9.
    Duff M.J.B: Review of the CLIP Image Processing System. AFIPS Conference Proceedings, Vol. 47, 1978.Google Scholar
  10. 10.
    Reddaway S.F: The DAP Approach. Infotech State of the Art Report on Supercomputers, Vol. 2, 1979.Google Scholar
  11. 11.
    Granlund G.H: In Search of a General Picture Processing Operator, Computer Graphics and Image Processing, Vol. 2, pp 155–173, 1978.Google Scholar
  12. 12.
    Granlund G.H: On One-dimensional Representation and Filtering of Image Information. Proc. International Conf. on Digital Signal Processing, Florence, Italy, 1978.Google Scholar
  13. 13.
    Granlund, G.H: An architecture of a Picture Processor Using a Parallel General Operator. Proc. from the Fourth International Joint Conference on Pattern Recognition, Kyoto, Japan, 1978.Google Scholar
  14. 14.
    Granlund G.H: Description of Texture Using the General Operator Approach. Proceedings of the 5th International Conference on Pattern Recognition, Miami, Florida, 1980.Google Scholar
  15. 15.
    Knutsson H, von Post B, and Granlund G.H: Optimization of Arithmetic Neighborhood Operations for Image Processing. Proceedings of the First Scandinavian Conference on Image Analysis Linköping, Sweden, 1980.Google Scholar
  16. 16.
    Knutsson H, and Granlund G.H: Fourier Domain Design of Line and Edge Detectors. Proceedings of the 5th International Conference on Pattern Recognition, Miami, Florida, 1980.Google Scholar
  17. 17.
    Brodatz P: Textures Dover, New York, 1966.Google Scholar
  18. 18.
    Rosenfeld A, and Thurston, M: Edge and Curve Detection for Visual Scene Analysis IEEE Tr on Computers, C-20, pp 562–569, 1971.Google Scholar
  19. 19.
    Granlund G.H: GOP, A Fast Parallel Processor for Image Information. Proceedings of the First European Signal Processing Conference Lausanne Switzerland, 1980.Google Scholar
  20. 20.
    Granlund G.H: GOP, A Fast and Flexible Processor for Image Analysis. Proceedings of the 5th International Conference on Pattern Recognition. Miami Beach, Florida, 1980.Google Scholar
  21. 21.
    Rosenfeld A, et al: Scene Labelling by Relaxation Operations IEEE Tr. Systems, Man & Cybernetics SMC-6, pp 420–433, 1976.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1981

Authors and Affiliations

  • Goesta H. Granlund
    • 1
  1. 1.Picture Processing LaboratoryLinkoeping UniversityLinkoepingSweden

Personalised recommendations