Mean Shift: A Robust Approach Toward Feature Space Analysis
IEEE Transactions on Pattern Analysis and Machine Intelligence
ICPR '02 Proceedings of the 16 th International Conference on Pattern Recognition (ICPR'02) Volume 4 - Volume 4
Utilization of multi-spectral images in photodynamic diagnosis
ICCVG'10 Proceedings of the 2010 international conference on Computer vision and graphics: Part II
High-throughput analysis of multispectral images of breast cancer tissue
IEEE Transactions on Image Processing
Hi-index | 0.00 |
This paper analyses multi-spectral images and their application in the process of cancer recognition in human skin. Cancerous part of a tissue can be characterized by higher accumulation of photosensitive substances then healthy. In order to detect the spectrum of Protoporphyrin IX in the human skin images Orthogonal Subspace Projection classifier was presented. For every pixel it calculates the content of Protoporphyrin IX spectrum in the global pixel spectrum. After pixel classification it was necessary to separate regions with cancer from healthy parts of a tissue by applying non-linear mapping with low frequency removal or mean shift segmentation enhanced with edge detection for better region recognition. Both proposals gave successful results.