Blind Correction of Human Lymphocyte Images by Optimal Thresholding in UWT Domain

Epidemiological studies have shown that the presence of chromosome damage or instability in human lymphocytes create the Micro Nucleuses (MNs) not incorporated into one of the daughter nuclei. Recently, the Image Flow Cytometer (IFC) was pointed out in order to acquire the image of the human lymphocytes, and to detect the MNs. The alterations introduced by the image acquisition system of the IFC causes errors in the automatic recognition, Therefore, proper methods were pointed out to correct them. The recurrent alterations are bad exposure, out of focus, motion blur, and Gaussian noise. In order to more increase the number of images correctly processed to detect and count the MNs, in the paper the new and optimized method able to perform the blind correction of the image affected by Gaussian noise is proposed. The method is based on the Un-decimated discrete Wavelet Transform (UWT). By taking into account the characteristics of the human lymphocyte images, the optimal value of threshold is evaluated with the aim to improve the quality of corrected image. Experimental tests are performed to compare the proposed correction method with the ones already available in the recent literature.