Regresar

Automatic estrogens detection in breast tumor images treated with immune-histochemical techniques

Abstract:

The work presented in this paper deals with medical imaging analysis. We have proposed an automatic and very fast algorithm to estimate estrogens presence in breast tumour tissues treated with immune-histochemical techniques. Colour image segmentation to isolate the different tissues has been performed in HSV, LUV and YIQ colour spaces. Thresholding techniques together with the euclidean distance algorithm are used. Subsequent refining processes are necessary to improve the accuracy of the estimate. Results are very close to those obtained by specialists in a manual way. This is part of a software tool for breast tumour detection, contributing to a higher level of assurance when estimating the seriousness of the tumour and its subsequent diagnosis.