A. K. Katsaggelos, "A multiple input image restoration approach," J. Vis. Commun. Image Representat., vol. 1, pp. 93--103, Sept. 1990.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....aim is the estimation of one or more restored image s given several correlated measured images. Such an application is relevant for color images where three (or more) measured spectral images are given [14] Another application is the restoration from several images with different exposure time [15, 16]. There are several differences between these applications and the super resolution ones. In the multiple image restoration, the resolution of the restored image is assumed the same as the one in the measurements, the resolution of all the measurements is the same, and the measured images are with ....

A. K. Katsaggelos, "A multiple input image restoration approach," J. of Visual Communication and Image Representation, vol. 1, pp. 93--103, September 1990.

....video is taken from the central loop. Since each of three prediction loops has its own reconstruction of the source frame, we can combine all three reconstructions to yield a better reconstruction than the one in the central loop. This is a special case of the multi channel restoration problem [4]. 3.1. Model The reconstructed video in each loop can be modeled as y i (l, m) x(l, m) n i (l, m) i = 1, 2, 3 (1) where i is the loop index, y i (l, m) is the reconstructed pixel value at position (l, m) in loop i, x(l, m) is the pixel at (l, m) in a source frame, and n i (l, m) is the ....

A. Katsaggelos, "A multiple input image restoration approach," Journal of Visual Communication and Image Representation, vol. 1, pp. 93--103, September 1990.

.... Bounding Ellipsoid Method A variant of the POCS based formulation using an ellipsoid to bound the constraint sets has been investigated by Tom and Katsaggelos [99, 100] and briefly mentioned by Elad and Feuer [85, 86] Given a set of ellipsoidal constraint sets, a bounding ellipsoid is computed [101]. The centroid of this ellipsoid is taken as the super resolution estimate. Since direct computation of this point is infeasible, an iterative procedure is used. It is interesting to note that this approach takes a form closely related to regularized methods. The observation model used in this ....

A. K. Katsaggelos, "A Multiple Input Image Restoration Approach," Journal of Visual Communication and Image Representation, vol. 1, no. 1, pp. 93--103, Sept. 1990.

....one [10] equivalent to the one presented in [5] Each scalar term within the summation in equation (2) represents an ellipsoid constraining the required superresolution ideal image, posed by the corresponding measured image. Since CLS is equivalent to the method of bounding ellipsoids [9], and the bounding ellipsoids method is closely related to the Projection Onto Convex Sets (POCS) method, our approach is similar to the one presented in [7] Since our suggested restoration process is iterative in nature, projections onto constraints represented by convex sets can be applied ....

A.K. Katsaggelos, "A Multiple Input Image Restoration Approach", J. of VCIR, vol. 1, pp. 93-103, September 1990.

....aim is the estimation of one or more restored image s given several correlated measured images. Such an application is relevant for color images where three (or more) measured spectral images are given [18] Another application is the restoration from several images with different exposure time [19, 20]. There are several differences between these applications and the super resolution ones. In the multiple image restoration, the resolution of the restored image is assumed the same as the one in the measurements, the resolution of all the measurements is the same, and the measured images are with ....

....result. 7 Summary and Conclusion A general new linear model has been developed in this work for the problem of superresolution restoration of continuous image sequence. This model is a generalization of the models used by the super resolution [14, 15, 16] and the multiple image restoration [20, 18, 19] algorithms. The principal model was shown to be the basis of the space and the frequency domains approaches for the problem of super resolution. Based on this model and the adaptive filtering theory [22, 21] several algorithms have been proposed, enabling the restoration of super resolution ....

[Article contains additional citation context not shown here]

A. K. Katsaggelos, "A multiple input image restoration approach," J. of Visual Communication and Image Representation, vol. 1, pp. 93--103, September 1990.

....a posteriori (MAP) estimation was presented, using a Gibbs prior. A linear model was used to describe the relationship between the low and high resolution images, and the motion was estimated by a block matching algorithm. In this paper, an iterative and temporally recursive technique based on [2] is used to improve the resolution of a video sequence. This paper is organized as follows. Section 2 presents the multiple input algorithm of [8] based on [2] Section 3 presents modifications to this algorithm which yield an improvement in the quality of the estimated high resolution image. ....

....and the motion was estimated by a block matching algorithm. In this paper, an iterative and temporally recursive technique based on [2] is used to improve the resolution of a video sequence. This paper is organized as follows. Section 2 presents the multiple input algorithm of [8] based on [2]. Section 3 presents modifications to this algorithm which yield an improvement in the quality of the estimated high resolution image. Experimental results are presented in Sec. 4, and conclusions are drawn in Sec. 5. 2. MULTIPLE INPUT ALGORITHM FOR IMAGE ENHANCEMENT In [2] an approach to ....

[Article contains additional citation context not shown here]

A. K. Katsaggelos, "A Multiple Input Image Restoration Approach," J. Visual Commun. Image Represent., vol. 1, pp. 93-103, Sept. 1990.

....frames which provides additional information about a given frame, enabling the increase of its spatial resolution. If the motion field is exactly known, then using the appropriate number of low resolution frames, a high resolution frame (and therefore sequence) can be exactly reconstructed [5]. Thus, the most important step in estimating high resolution sequences is that of motion estimation. However, it is well known that motion estimation is a very difficult problem due to 1) its ill posedness [6] 2) the aperture problem, and 3) the presence of covered and uncovered regions. A ....

....The main reason for this approach is that a simple optimality criterion such as the mean absolute error (MAE) may not yield the most visually appealing result. In this paper an iterative video resolution enhancement algorithm is proposed, based on the multiple input restoration algorithm of [5]. The low resolution frames, appropriately compensated for the motion and upsampled, represent the multiple observations of a high resolution frame. The iterative algorithm recovers the high resolution frame from these multiple degraded versions. Various ways to estimate the motion and to ....

[Article contains additional citation context not shown here]

A. K. Katsaggelos, "A multiple input image restoration approach," J. Vis. Commun. Image Represent., vol. 1, pp. 93--103, Sept. 1990.

....been demonstrated in [8, 9] In both of these approaches, the frames surrounding the current working frame are first motion compensated to the current frame. In [8] a non recursive approach using maximum a posteriori (MAP) estimation is presented, using a Gibbs prior. In [9] a technique based on [6] is used to improve the resolution of a video sequence. However, work in color video processing has been significantly more limited. In this paper, the extension of [9] to color video resolution enhancement will be discussed. From [7] it is recognized that the motion estimation of color sequences ....

....the spatial information contained in the adjacent frames. These adjacent frames need to first be motion compensated to the current frame. In the case of perfect motion compensation, each frame in the sequence can be considered to be identical to the working frame. This is the assumption made in [6], where multiple distorted versions of the same original image are available. This paper is organized as follows. In Sec. 2, notation is introduced and the iterative algorithm is presented. Section 3 discusses the importance of the motion estimator, and presents two approaches for motion ....

[Article contains additional citation context not shown here]

A. K. Katsaggelos, "A Multiple Input Image Restoration Approach," J. Visual Commun. Image Represent., vol. 1, pp. 93-103, Sept. 1990.

No context found.

A.K. Katsaggelos, "A Multiple Input Image Restoration Approach," J. Visual Commun. Image Represent., vol. 1, pp. 93-103, Sept. 1990.

No context found.

A. K. Katsaggelos, "A multiple input image restoration approach," J. Vis. Commun. Image Representat., vol. 1, pp. 93--103, Sept. 1990.

No context found.

A. K. Katsaggelos, "A multiple input image restoration approach," J. Vis. Commun. Image Representat., vol. 1, pp. 93--103, Sept. 1990.

No context found.

A. Katsaggelos, "A multiple input image restoration approach," Journal of Visual Communication and Image Representation, vol. 1, no. 1, pp. 93--103, September 1990.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC