Please use this identifier to cite or link to this item:
http://acervodigital.unesp.br/handle/11449/21795
- Title:
- Multiscale Fractal Descriptors and Polynomial Classifier for Partial Pixels Identification in Regions of Interest of Mammographic Images
- Universidade Federal de Uberlândia (UFU)
- IFTM
- Universidade Estadual Paulista (UNESP)
- Universidade Federal do ABC (UFABC)
- Faculdade de Medicina de São José do Rio Preto (FAMERP)
- 1548-0992
- Computer systems are used to support breast cancer diagnosis, with decisions taken from measurements carried out in regions of interest (ROIs). We show that support decisions obtained from square or rectangular ROIs can to include background regions with different behavior of healthy or diseased tissues. In this study, the background regions were identified as Partial Pixels (PP), obtained with a multilevel method of segmentation based on maximum entropy. The behaviors of healthy, diseased and partial tissues were quantified by fractal dimension and multiscale lacunarity, calculated through signatures of textures. The separability of groups was achieved using a polynomial classifier. The polynomials have powerful approximation properties as classifiers to treat characteristics linearly separable or not. This proposed method allowed quantifying the ROIs investigated and demonstrated that different behaviors are obtained, with distinctions of 90% for images obtained in the Cranio-caudal (CC) and Mediolateral Oblique (MLO) views.
- 1-Jun-2012
- IEEE Latin America Transactions. Piscataway: IEEE-Inst Electrical Electronics Engineers Inc, v. 10, n. 4, p. 1999-2005, 2012.
- 1999-2005
- Institute of Electrical and Electronics Engineers (IEEE)
- Mammography
- Regions of Interest
- Partial Pixels
- Fractal Descriptors
- Polynomial Classifier
- http://dx.doi.org/10.1109/TLA.2012.6272486
- Acesso restrito
- outro
- http://repositorio.unesp.br/handle/11449/21795
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.