Combined local color and texture analysis of stained cells
Computer Vision, Graphics, and Image Processing
Fractals everywhere
Texture Measures for Automatic Classification of Pulmonary Disease
IEEE Transactions on Computers
Hi-index | 0.00 |
An important tool in medical diagnosis is digital images analysis and interpretation. The objective of the present paper is to demonstrate the utility of fractal analysis in theoretic decision for doubtful cases in malignity of mammary tumours. We present the fractal and textural features able to classify these tumours in malign or benign. Thus, the most important feature is the local connected fractal dimension calculated for edges extracted from binary images obtained with different binary thresholds. We describe a tumour classification algorithm based on the histogram of the local connected fractal dimension. In order to validate the described algorithm, experimental results from 30 images (mammography) are presented. At the work ending, some concluding remarks are presented also.