Digital mammograms analysis using fractal and multifractal methods
| Autor: | Nadia Kermouni Serradj, Mohammed Messadi, Sihem Lazzouni |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
Holder exponent
business.industry Computer science Quantitative Biology::Tissues and Organs Physics::Medical Physics Pattern recognition 02 engineering and technology Multifractal system 010402 general chemistry 021001 nanoscience & nanotechnology 01 natural sciences Fractal dimension 0104 chemical sciences Box counting Fractal Computer Science::Computer Vision and Pattern Recognition Lacunarity Segmentation Artificial intelligence 0210 nano-technology business |
| Zdroj: | 2019 6th International Conference on Image and Signal Processing and their Applications (ISPA). |
| DOI: | 10.1109/ispa48434.2019.8966813 |
| Popis: | This paper presents a fractal and multifractal analysis for detection and segmentation of breast abnormalities using box counting method. The first step is the detection of the pathological breast by comparing parameters extracted from the multifractal spectrum of pairs of mammograms. The second consists to find the suspect region by decomposing the pathological mammogram into four blocks. The suspect region is characterized by a low value of Holder exponent, large values of asymmetry, lacunarity and fractal dimension. Finally, segmentation of some abnormalities is made on α-image and compared with k-means method. The whole process was applied on images of the Mini-Mias database. The obtained results are promising. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|