Back to Search
Start Over
Acute lymphoblastic leukemia image segmentation driven by stochastic fractal search
- Source :
- Multimedia Tools and Applications. 79:12227-12255
- Publication Year :
- 2020
- Publisher :
- Springer Science and Business Media LLC, 2020.
-
Abstract
- Cancer is one of the most critical disease. In particular, Leukemia is the most common type of cancer which produces an excessive quantity of leucocytes, replacing normal blood cells. Early detection of leucocytes cells can save human life. Recently, researchers have contributed to the development of computer assisted pathology techniques to automatically detect cancer at early stage. Commonly, assisted pathology systems are based on artificial vision techniques to identify cancer cells in the human body. Blood image segmentation techniques for Leukemia have been proposed based on automatic thresholding schemes involving traditional clustering methods. However, traditional clustering methods are sensitive to initial cluster positions, where the incorrect centering values results into false positive cancer diagnosis. On the other hand, Nature-Inspired Optimization Algorithms (NIOA) are stochastic search methods for finding the optimal solution for complex multimodal functions where traditional optimization approaches are not suitable to operate. Since blood image segmentation is considered as a complex computational task, NIOA methods yield an interesting alternative to proper blood cell segmentation. In this paper, the Stochastic Fractal Search (SFS) algorithm is implemented in order to provide non-false positive segmented outcomes for Leukemia identification. In the experimental study, the proposed approach is compared against traditional clustering methods as well as some NIOAs techniques. The numerical results indicate that SFS, provide superior results in terms of accuracy, time complexity, and quality parameters.
- Subjects :
- Computer Networks and Communications
Computer science
business.industry
Cancer
Pattern recognition
Image segmentation
medicine.disease
Thresholding
Identification (information)
Fractal
Hardware and Architecture
Media Technology
medicine
Segmentation
Artificial intelligence
business
Cluster analysis
Time complexity
Software
Subjects
Details
- ISSN :
- 15737721 and 13807501
- Volume :
- 79
- Database :
- OpenAIRE
- Journal :
- Multimedia Tools and Applications
- Accession number :
- edsair.doi...........0afdfa29ec130b6584676bcfdfbb9c75