Fractal analysis and fuzzy c-means clustering for quantification of fibrotic microscopy images

The advances in improved fluorescent probes and better cameras in collaboration with the advent of computers in imaging and image analysis, assist the task of diagnosis in microscopy imaging. Based on such technologies, we introduce a computer-assisted image characterization tool based on fractal analysis and fuzzy clustering for the quantification of degree of the Idiopathic Pulmonary Fibrosis in microscopy images. The implementation of this algorithmic strategy proved very promising concerning the issue of the automated assessment of microscopy images of lung fibrotic regions against conventional classification methods that require training such as neural networks. Fractal dimension is an important image feature that can be associated with pathological fibrotic structures as is shown by our experimental results.