Extracting feature vectors of biomedical images

We present few approaches for extracting feature vectors from biomedical images with cervical cancer. In the first case we use the technique of statistical wavelet analysis. In the second case we use wavelet analysis in "Brute force" approach. In the third case we combine two above approaches. On the first step an image is subjected to discrete wavelet transformation. On the following step the wavelet analysis of transformation coefficients is carried out. After extracting of feature vectors we used neural net to classify biomedical images. Also we have compared approaches to find best feature vectors.