Image denoising and image enhancement on the applications of confocal laser scanning microscopy

Confocal laser scanning microscopy (CLSM) is a developing optical imaging device enabling non-invasive examination of live biological tissues with laser light in realtime. CLSM provides optical sectioning of samples. Image can get corrupted with noise of different levels due to out-of-focus light back-scattered above and below the focal plane. Construction of the CLSM setup is established and several images are captured. This work attempts to analyze the effects of different denoising and contrast enhancement techniques by using real CLSM images with the help of different image quality metrics. Additive white Gaussian noise (AWGN) is used as a noise model. A reliable method for estimating the standard deviation of AWGN in a single image is also performed on real CLSM images. Wavelet transform is the most effective candidate for noise suppression since it is capable of preserving energy conservation during inverse transformation. A denoising algorithm is developed to make it applicable on CLSM. An important issue that affects the performance of 2D-DWT is the selection of components employed in the algorithm along with their parameter selection. This study examines the effect of employing different combinations of 2D-DWT components and tuning parameter values on different image quality assessments. Design of Experiments (DOE) is presented as a systematic approach to catch the best combination of these parameter values. Analysis of variance (ANOVA) is used to inspect the main effect and interaction effects of the treated parameters. Computational results verified the efficacy of the proposed algorithm and the methodical approach for the image denoising of CLSM images. After denoising, several histogram equalization methods are put into practice for contrast enhancement. The comparison of methods that give better enhancement result is provided with the means of different quantative measures for better visualization.