Image Enhancement Based on Fractional Poisson for Segmentation of Skin Lesions Using the Watershed Transform
Image segmentation is considered as a necessary step towards accurate medical analysis by extracting the crucial medical information in identifying abnormalities. This study proposes a new technique for segmentation a malignant melanoma in images. A new filter is proposed for smoothing input images and more accurate segmentation based on fractional Poisson. In the pre-processing step, eight masks of size n × n are created to eliminate noise and obtain a smooth image. The watershed algorithm is used for segmentation with morphological operation to better segment the skin lesion area. The proposed method was capable of improving the accuracy of the segmentation up to 96.47%.
KeywordsFractional calculus Fractional Poisson Watershed algorithm Skin lesion Segmentation
The authors would like to thank both Dr. Joaquim M. da Cunha Viana and Mr. Navid Razmjooy for providing the dermoscopic images used in this study. We would also like to thank skin specialist Dr. Mohammed Ahmed, for providing the necessary information for this study. This research is supported by the Fundamental Research Grant Scheme (FRGS), Project: FP073-2015A from Ministry of Higher Education, Malaysia.
- 2.Wu, Q., Merchant, F., Castleman, K.: Microscope Image Processing. Academic Press, Cambridge (2010)Google Scholar
- 4.Jung, C.R.: Multiscale image segmentation using wavelets and watersheds. In: 2003 XVI Brazilian Symposium on Computer Graphics and Image Processing, SIBGRAPI 2003, pp. 278–284. IEEE (2003)Google Scholar
- 7.Ahmed Abbas, A., Tan, W.-H., Guo, X.-N.: Combined optimal wavelet filters with morphological watershed transform for the segmentation of dermoscopic skin lesions. In: Anthony, P., Ishizuka, M., Lukose, D. (eds.) PRICAI 2012. LNCS, vol. 7458, pp. 722–727. Springer, Heidelberg (2012). doi: 10.1007/978-3-642-32695-0_63 CrossRefGoogle Scholar
- 10.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.: Modified watershed technique and post-processing for segmentation of skin lesions in dermoscopy images. Comput. Med. Imaging Graph. 35, 116–120 (2011)CrossRefGoogle Scholar
- 16.Stoev, S.L.: A fast watershed algorithm based on rainfalling simulation (2000)Google Scholar
- 17.Sarker, M.S.Z., Haw, T.W., Logeswaran, R.: Morphological based technique for image segmentation. Int. J. Inf. Technol. 14, 55–80 (2008)Google Scholar
- 20.Wu, Y., Xie, F., Jiang, Z., Meng, R.: Automatic skin lesion segmentation based on supervised learning. In: 2013 Seventh International Conference on Image and Graphics (ICIG), pp. 164–169. IEEE (2013)Google Scholar
- 23.Cavalcanti, P.G., Yari, Y., Scharcanski, J.: Pigmented skin lesion segmentation on macroscopic images. In: 2010 25th International Conference of Image and Vision Computing New Zealand (IVCNZ), pp. 1–7. IEEE (2010)Google Scholar