All Figures

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg1.eps} \end{figure}$ Figure 1: One- and two-dimensional bases for the computation of the masks using the linear interpolation approach. In the one-dimensional case the dashed line indicates the limits of each subregion, whereas in the two-dimensional case they indicate the limits of the masks. Symbols $\times$ denote the center of a subregion, whereas the values of the masks at the marked points are given by $\bullet = 1$ , $\blacklozenge =0.5$ , $\blacktriangle =0.25$ , $\blacksquare =0$ .
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=15cm,clip]{0754Fg2.eps} \end{figure}$ Figure 2: Spatial variation of the PSF across the first image of the set used in the numerical experiments (see text). The pair of numbers in the header of each panel indicates the coordinates of the pixels to which the displayed PSF corresponds. In the intermediate pixels the PSFs are obtained through a linear interpolation of the displayed ones. Each PSF is given by a two-dimensional Gaussian function with a dispersion of 12 pixels along the major axis and 4 pixels along the minor axis.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg3.eps} \end{figure}$ Figure 3: Original image, blurred version and blurred + noise version of the first image of the set (see text), and corresponding mean PSF. The images are $400 \times 400$ pixels.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg4.eps} \end{figure}$ Figure 4: Mean image $\zeta (m,n)$ (with D_i(m,n) obtained using piecewise constant interpolation), MRNSD and Tikhonov deblurred versions of $\zeta (m,n)$ , and MRNSD deblurred version of the first image of the set (see text) for the image shown in Fig. 3. For the MRNSD method the images shown are those obtained after 300 iterations of the algorithm, whereas for Tikhonov it is the one with the smallest rms of the true residuals; for the mean image and first image of the set the rms of the true residuals are, respectively, 0.028 (MRNSD) - 0.032 (Tikhonov), and 0.032.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg5.eps} \end{figure}$ Figure 5: Original image, blurred version and blurred + noise version of the first image of the set (see text), and corresponding mean PSF. The images are $400 \times 400$ pixels.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg6.eps} \end{figure}$ Figure 6: Mean image $\zeta (m,n)$ (with D_i(m,n) obtained using piecewise constant interpolation), MRNSD and Tikhonov deblurred versions of $\zeta (m,n)$ , and MRNSD deblurred version of the first image of the set (see text) for the image shown in Fig. 5. The images shown are the ones with the smallest rms of the true residuals; for the mean image and first image of the set these are, respectively, 0.082 (MRNSD) - 0.088 (Tikhonov), and 0.109.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg7.eps} \end{figure}$ Figure 7: Mean image $\zeta (m,n)$ (with D_i(m,n) obtained using linear interpolation), MRNSD and Tikhonov deblurred versions of $\zeta (m,n)$ , and MRNSD deblurred version of the first image of the set (see text) for the image shown in Fig. 5. The images shown are the ones with the smallest standard deviation of the true residuals; for the mean image and first image of the set these are, respectively, 0.076 (MRNSD) - 0.082 (Tikhonov) and 0.109.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg8.eps} \end{figure}$ Figure 8: Original image, blurred version and blurred + noise version of the first image of the set (see text), and corresponding mean PSF. The images are $400 \times 400$ pixels, but only the central $320 \times 320$ pixels are shown (see text).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg9.eps} \end{figure}$ Figure 9: Mean image $\zeta (m,n)$ (with D_i(m,n) obtained using linear interpolation), MRNSD and Tikhonov deblurred versions of $\zeta (m,n)$ , and MRNSD deblurred version of the first image of the set (see text) for the image shown in Fig. 8. The images shown are the ones with the smallest rms of the true residuals; for the mean image and first image of the set these are, respectively, 0.036 (MRNSD) - 0.036 (Tikhonov), and 0.043.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg10.eps} \end{figure}$ Figure 10: Mean image $\zeta (m,n)$ (with D_i(m,n) obtained using linear interpolation), MRNSD and Tikhonov deblurred versions of $\zeta (m,n)$ , and MRNSD deblurred version of the first image of the set (see text) for the image shown in Fig. 5 and Poissonian noise (peak signal to noise ratio = $40~{\rm dB}$ ). The images shown are the ones with the smallest rms of the true residuals; for the mean image and first image of the set these are, respectively, 0.076 (MRNSD) - 0.082 (Tikhonov) and 0.109.
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg11.eps} \end{figure}$ Figure 11: Mean image $\zeta (m,n)$ (with D_i(m,n) obtained using linear interpolation), MRNSD and Tikhonov deblurred versions of $\zeta (m,n)$ , and MRNSD deblurred version of the first image of the set (see text) for the image shown in Fig. 8 and Poissonian noise (peak signal to noise ratio = $40~{\rm dB}$ ). The images shown are the ones with the smallest rms of the true residuals; for the mean image and first image of the set these are, respectively, 0.037 (MRNSD) - 0.037 (Tikhonov), and 0.044 (NB. here the values of the pixels have been converted into photon counts).
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg12.eps} \end{figure}$ Figure 12: Original image, observed image at $100~{\rm GHz}$ , mean image $\zeta (m,n)$ (with D_i(m,n) obtained using linear interpolation), and Tikhonov deblurred versions of $\zeta (m,n)$ .
Open with DEXTER

In the text

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg3.eps} \end{figure}$	Figure 3: Original image, blurred version and blurred + noise version of the first image of the set (see text), and corresponding mean PSF. The images are $400 \times 400$ pixels.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg5.eps} \end{figure}$	Figure 5: Original image, blurred version and blurred + noise version of the first image of the set (see text), and corresponding mean PSF. The images are $400 \times 400$ pixels.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg8.eps} \end{figure}$	Figure 8: Original image, blurred version and blurred + noise version of the first image of the set (see text), and corresponding mean PSF. The images are $400 \times 400$ pixels, but only the central $320 \times 320$ pixels are shown (see text).
Open with DEXTER

$\begin{figure} \par\includegraphics[width=14cm,clip]{0754Fg12.eps} \end{figure}$	Figure 12: Original image, observed image at $100~{\rm GHz}$ , mean image $\zeta (m,n)$ (with D_i(m,n) obtained using linear interpolation), and Tikhonov deblurred versions of $\zeta (m,n)$ .
Open with DEXTER