Improvement in volume estimation from confocal sections after image deconvolution

The confocal microscope can image a specimen in its natural environment forming a 3D image of the whole structure by scanning it and collecting light through a small aperture (pinhole), allowing in vivo and in vitro observations. So far, the confocal fluorescence microscope (CFM) is considered a true volume imager because of the role of the pinhole that rejects information coming from out-of-focus planes. Unfortunately, intrinsic imaging properties of the optical scheme presently employed yield a corrupted image that can hamper quantitative analysis of successive image planes. By a post-image collection restoration, it is possible to obtain an estimate, with respect to a given optimization criterium, of the true object, utilizing the impulse response of system or Point Spread Function (PSF). The PSF can be measured or predicted so as to have a mathematical and physical model of the image-formation process. Further modelling and recording noise as an additive Gaussian process has used the regularized Iterative Constrained Tykhonov Miller (ICTM) restoration algorithm for solving the inverse problem. This algorithm finds the best estimate iteratively searching among the possible positive solutions; in the Fourier domain, such an approach is relatively fast and elegant. In order to compare the effective improvement in the quantitative image information analysis, we measured the volume of reference objects before and after image restoration, using the isotropic Fakir method. (C) 2004 Wiley-Liss, Inc.