The Pruning of Combined Neighborhood Differences Texture Descriptor for Multispectral Image Segmentation
This paper proposes a novel feature extraction method for unsupervised multispectral image segmentation by pruning the two dimensional texture feature named combine neighborhood differences. In contrast with the original CND, which is applied with traditional image, the pruned CND is computed o...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
2011
|
| Subjects: | |
| Online Access: | http://eprints.utem.edu.my/id/eprint/4103/1/IJCTEE_1111_26.pdf http://eprints.utem.edu.my/id/eprint/4103/ http://www.ijctee.org/ISSUE3.html |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Universiti Teknikal Malaysia Melaka |
| Language: | English |
| Summary: | This paper proposes a novel feature extraction
method for unsupervised multispectral image segmentation by
pruning the two dimensional texture feature named combine
neighborhood differences. In contrast with the original CND,
which is applied with traditional image, the pruned CND is
computed on a single pixel with various bands. The proposed
algorithm utilizes the sign of differences between bands of the
pixel. The difference values are thresholded to form a binary
codeword. A binomial factor is assigned to the codeword to form
another unique value. These values are then grouped to
construct the multiband CND feature image where is used in the
unsupervised segmentation. Experimental results, with respect
to segmentation and classification accuracy using two
LANDSAT multispectral images test suite have been performed.
The result shows that the pruned CND feature outperforms
spectral feature, with average classification accuracies of
87.55% whereas that of spectral feature is 55.81%. |
|---|