8.3.2 Complete Systems Derived from the Univ. Massachusetts Work

Chapter Contents (Back)
Segmentation, Histogram. System: VISIONS.
See also University of Massachusetts VISIONS System.

Beveridge, J.R.[J. Ross], Griffith, J.S.[Joey S.], Kohler, R.R.[Ralf R.], Hanson, A.R., and Riseman, E.M.,
Segmenting Images Using Localized Histograms and Region Merging,
IJCV(2), No. 3, January? 1989, pp. 311-352.
Springer DOI BibRef 8901
And: COINSTR 87-88. October 1987. Segmentation, Systems. Segmentation, Region Merging. This is the newest UMass segmentation ideas. The first step is to divide the region into overlapping sectors and generate thresholds based on their histograms. Adjacent thresholds are propagated if necessary and these are merged. The results are good, but the complexity is high. This uses a lot of the ideas from the earlier segmentation work and changes things since the earlier ideas did not work that well. BibRef

Hanson, A.R., and Riseman, E.M.,
Segmentation of Natural Scenes,
CVS78(xx-yy). System: VISIONS. BibRef 7800

Nagin, P.A., Hanson, A.R., and Riseman, E.M.,
Studies in Global and Local Histogram Guided Relaxation Algorithms,
PAMI(4), No. 3, May 1982, pp. 263-277. BibRef 8205
Region Extraction and Description Through Planning,
COINS-TR 77-8, May 1977. BibRef
Earlier: A3, A2, A1:
Region Growing in Textured Outdoor Scenes,
UMass-TR-75C-3, February 1975. Segmentation, Systems. Segmentation. Relaxation. Histogram guided is limited to initial threshold selection based on significant peaks, which, when there is substantial overlap, results in mixed assignments and in breaking up regions. All assignments are made all at once. (n-tuple of probability values) Relax to up date assignment vectors. Suppress very small regions. Checks each region (multi-model) individually. No references to Ohlander! A set of artificial image examples which should be trivial using OPR-thing diagonal and lines may be hard. They also introduce local segmentation (several (4) parts of the image and combine results at borders, also use local results to improve compatibility coefficients. BibRef

Price, K.E.,
Image Segmentation: A Comment on 'Studies in Global and Local Histogram-Guided Relaxation Algorithms',
PAMI(6), No. 2, March 1984, pp. 247-249. BibRef 8403 USC Computer VisionDiscussion of some of the flaws in the named paper. BibRef

Hanson, A.R., Riseman, E.M., and Nagin, P.A.,
Authors' Reply,
PAMI(6), No. 2, March 1984, pp. 249. BibRef 8403

Kohler, R.R.[Ralf R.],
A Segmentation System Based on Thresholding,
CGIP(15), No. 4, April 1981, pp. 319-338.
Elsevier DOI Segmentation, Thresholds. Segmentation, Edges. Edges which correspond to real boundaries tend to have high contrast, so the optimum threshold is the one that detects more high contrast edges and fewer low contrast edges. BibRef 8104

Kohler, R.R.[Ralf R.],
Integrating Non-Semantic Knowledge into Image Segmentation Processes,
Ph.D.Thesis (CS), 1984, BibRef 8400 COINS-TR-84-04, UMass. Segmentation, Knowledge. Also discusses systems issues for linking Fortran and Lisp. BibRef

Nagin, P.A.,
Studies in Image Segmentation Algorithms Based on Histogram Clustering and Relaxation,
Ph.D.Thesis (CS), September 1979. BibRef 7909 COINS-TR-79-15, UMass..
See also Studies in Global and Local Histogram Guided Relaxation Algorithms. for paper from this. BibRef

Nagin, P.A., Kohler, R.R., Hanson, A.R., and Riseman, E.M.,
Segmentation, Evaluation, and Natural Scenes,
PRIP79(515-522). BibRef 7900

Prager, J.M.,
Extracting and Labeling Boundary Segments in Natural Scenes,
PAMI(2), No. 1, January 1980, pp. 16-26. BibRef 8001
Earlier: Add: A2, A3: Hanson, A.R., and Riseman, E.M., COINSTR 77-7, May 1977. Relaxation. Determine whether the segments are menaingful. BibRef

Prager, J.M.,
Segmentation of Static and Dynamic Scenes,
MITCS TR 79-7, May 1979. BibRef 7905

Prager, J.M., Nagin, P.A., Kohler, R.R., Hanson, A.R., and Riseman, E.M.,
Segmentation Processes in the VISIONS System,
IJCAI77(642-644). BibRef 7700

Chapter on 2-D Region Segmentation Techniques, Snakes, Active Contours continues in
Segmentation by Thresholding, Quantization, or Relaxation .

Last update:May 6, 2024 at 15:50:14