Modified watershed technique and post-processing for segmentation of skin lesions in dermoscopy images Journal Article
| Authors: | Wang, H.; Moss, R. H.; Chen, X.; Stanley, R. J.; Stoecker, W. V.; Celebi, M. E.; Malters, J. M.; Grichnik, J. M.; Marghoob, A. A.; Rabinovitz, H. S.; Menzies, S. W.; Szalapski, T. M. |
| Article Title: | Modified watershed technique and post-processing for segmentation of skin lesions in dermoscopy images |
| Abstract: | In previous research, a watershed-based algorithm was shown to be useful for automatic lesion segmentation in dermoscopy images, and was tested on a set of 100 benign and malignant melanoma images with the average of three sets of dermatologist-drawn borders used as the ground truth, resulting in an overall error of 15.98%. In this study, to reduce the border detection errors, a neural network classifier was utilized to improve the first-pass watershed segmentation; a novel "edge object value (EOV) threshold" method was used to remove large light blobs near the lesion boundary; and a noise removal procedure was applied to reduce the peninsula-shaped false-positive areas. As a result, an overall error of 11.09% was achieved. © 2010 Elsevier Ltd. |
| Keywords: | malignant melanoma; segmentation; image processing; watershed; neural network |
| Journal Title: | Computerized Medical Imaging and Graphics |
| Volume: | 35 |
| Issue: | 2 |
| ISSN: | 0895-6111 |
| Publisher: | Elsevier Inc. |
| Date Published: | 2011-03-01 |
| Start Page: | 116 |
| End Page: | 120 |
| Language: | English |
| DOI: | 10.1016/j.compmedimag.2010.09.006 |
| PROVIDER: | scopus |
| PUBMED: | 20970307 |
| PMCID: | PMC3183575 |
| DOI/URL: | |
| Notes: | --- - "Export Date: 4 March 2011" - "CODEN: CMIGE" - "Source: Scopus" |
