Name
Abstract | ||
|---|---|---|
In this paper we present a new postal envelope segmentationmethod based on 2-D histogram clustering and watershedtransform. Segmentation task consists in detectingthe modes associated with homogeneous regions in envelopeimages such as handwritten address block, postmarks,stamps and background. The homogeneous modes in 2-Dhistogram are segmented through the morphological watershedtransform. Our approach is applied to complex Brazilianpostal envelopes. Very little a priori knowledge of theenvelope images is required. The advantages of this approachwill be described and illustrated with tests carriedout on 300 different images where there are no fixed positionfor the handwritten address block, postmarks and stamps. |
Year | DOI | Venue |
|---|---|---|
2003 | 10.1109/ICDAR.2003.1227685 | ICDAR-1 |
Keywords | Field | DocType |
wavelet transforms,image recognition,a priori knowledge,watershed transform | Computer vision,Histogram,Pattern recognition,Segmentation,Computer science,Pattern clustering,Homogeneous,A priori and a posteriori,Watershed,Artificial intelligence,Cluster analysis,Wavelet transform | Conference |
ISSN | ISBN | Citations |
1520-5363 | 0-7695-1960-1 | 5 |
PageRank | References | Authors |
0.61 | 10 | 2 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Eduardo Akira Yonekura | 1 | 9 | 1.50 |
| Jacques Facon | 2 | 67 | 15.67 |