Measure of image clarity using image features extracted by the multiscale top-hat transform

To correctly quantify the clarity of an image and achieve good discrimination, an effective measure of image clarity using the image features extracted by the multiscale top-hat transform is presented in this paper. Firstly, multiscale image features are extracted by the top-hat transform using multiscale structuring elements. Then, the final extracted image features for constructing the measure of image clarity are obtained from the extracted multiscale image features. Finally, the mean value of the final extracted image features is calculated as the proposed measure of image clarity. Comparison experiments and application to image clarity enhancement of mineral images show that the proposed measure is effective for measuring image clarity and has good discrimination ability compared with some other measures. Therefore, the proposed image clarity measure may be used for different applications such as automatic focusing in microscopy, image enhancement, image fusion and so on. (paper)