ABSTRACT
Presently, advantages of digital image processing and analysis are recognizable in all areas, including pathology. Many digital signal-processing techniques have been successfully developed in the last decades. However, for recognizing and understanding complex cellular patterns in light microscope images, a few automated approaches were proposed. Most of the proposed methods are based on the machine vision system, both supervised and unsupervised, and wavelet feature selection. Wavelet technique can handle complex patterns, and the available features use the discriminative properties of the cells for identification and classification purposes. This chapter describes specific procedures for wavelet transformations such as de-noising, image enhancement, edge detection, and edge preserved through image fusion used for digital image processing, with specific reference to fluorescence microscopy images. Microscopy digital image-processing techniques are used for both enhancement and manipulation of raw images to make the informational content available. Microscopy digital image analysis techniques comprehend provided data by mean of data recognition techniques or image-quality assessment. Wavelet decomposition is performed on the cell images and a fusion technique related to horizontal and vertical orientations is implemented for best edge-preserving results. The Haar wavelet transform has been used to get the approximation and detail coefficients at 1- and 2-level of decomposition and, further, as a tool for image processing (such as de-noising, enhancing, edge detection, and edge preserving) and image analysis (such as image-quality assessment). Correlation between the results from the machine vision system (wavelet decomposition, de-noising, edge map, and fusion) and quality metrics is improved. To make the subject accessible and practical, a number of applications are provided. Moreover, the code listing for each application is provided. The MATLAB software is used to identify and choose the adequate solution for a particular problem.
