ENHANCED GRAPH BASED NORMALIZED CUT METHODS FOR IMAGE SEGMENTATION
| Autor: | Kapade S D, Chaudhari B S, Khairnar S M |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2014 |
| Předmět: |
Multiscale
Normalized Cut Computer science Segmentation-based object categorization business.industry Graph based ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Scale-space segmentation Pattern recognition Image segmentation Image Segmentation lcsh:Computer applications to medicine. Medical informatics lcsh:Telecommunication Pixel Affinity lcsh:TK5101-6720 lcsh:R858-859.7 Artificial intelligence Watershed Regions business Connected-component labeling |
| Zdroj: | ICTACT Journal on Image and Video Processing, Vol 5, Iss 2, Pp 907-911 (2014) |
| ISSN: | 0976-9102 0976-9099 |
| Popis: | Image segmentation is one of the important steps in digital image processing. Several algorithms are available for segmenting the images, posing many challenges such as precise criteria and efficient computations. Most of the graph based methods used for segmentation depend on local properties of graphs without considering global impressions of image, which ultimately limits segmentation quality. In this paper, we propose an enhanced graph based normalized cut method for extracting global impression and consistencies in the image. We propose a technique to add flexibility to original recursive normalized two way cut method which was further extended to other graph based methods. The results show that the proposed technique improves segmentation quality as well as requires lesser computational time than the regular normalized cut method. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|