S. Alliney and C. Morandi, "Digital image registration using projections," IEEE Trans. Pattern Anal. Machine Intell., vol. 8, no. 2, pp. 222--233, Mar. 1986.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....of an image is defined as line integrals across the image [10] It is well known that pure translational motion in an image results in translation of the projections [10] along the direction of projection. This property has been used successfully in the past to estimate motion using projections [1, 2, 8, 9, 15, 18, 20 22, 24 26]. More recently, we have unified much of the (mostly ad hoc) work in this area and proposed a model of more general motion vector fields in the Radon transform domain [19] In particular, it can be shown, as will be elaborated below, that affine motion in the image leads to affine motion in the ....

....g(p,#) g( p,# # ) We note here that while the above definition represents the model for the Radon transform of a continuous image, we will in practice use a discrete version of the Radon transform. The use of projections to efficiently estimate motion is not new. Very early works such as [2] use image projections at 0# and 90# to register translated images using a relative phase approach. More recently [15, 25] have incorporated projections into correlation based block motion estimators to speed up motion compensated video coding. In these works, the projections used to estimate ....

S. Alliney and C. Morandi, "Digital image registration using projections," IEEE Trans. Pattern Anal. Machine Intell., Vol. 8, No. 2, pp. 222--233, 1986.

.... [11] By computing the phase correlation via the Fast Fourier Transform, robust results are achieved, which can be implemented on real time hardware [16] Faster computation can be achieved at the cost of robustness and accuracy if the phase correlation is computed only with one dimensional FFTs [1]. 7] and [12] propose frequency domain algorithms that can search for rotation and affine transformations respectively, but they are computationally expensive. Figure 2: Mapping to a 2 D World Model However, the image rotation (about the optical axis) and scaling information of an image can be ....

S. Alliney and C. Morandi, "Digital Image Registration Using Projections," IEEE Transactions on PAMI, Vol. 8, No. 2, March 1986, pp. 222-233.

....thus increasing the resolution and improving the approximation of the transform after rotation. Other interpolation techniques, for instance, nearest neighbor and bilinear interpolation, proved to be unsatisfactory. Their method is also costly because of the difficulty in testing for each OE. Alliney 86] presented a method which only requires one dimensional Fourier transformations to compute the phase correlation. By using the xand y projections of each image, the Fourier transforms are given by the projection slice theorem. The 1D transforms of the x and y projections are simply the row of ....

S. Alliney and C. Morandi, "Digital Image Registration using Projections," IEEE Trans. Pattern Analyis and Machine Intelligence PAMI-8 , No. 2, March 1986, pp222-233.

....of estimating this offset at subpixel accuracy, and using this estimate to register the second image to the grid of the first image. This problem has been attacked in the signal or pixel domain [1, 6, 7, 9, 10, 11, 14, 15, 16, 17, 18, 19, 20, 22, 24, 23, 27, 30, 31, 33] and in the Fourier domain [2, 3, 4, 5, 7, 8, 12, 13, 21, 25, 26, 28, 29, 32]. Our work follows the latter body, and estimates the shift from basic phase relationships between the Fourier transform of the two images. However, unlike previous work, we do not assume that each observed image represents alias free samples of an underlying continuous image. In fact, point ....

S. Alliney and C. Morandi. Digital image registration using projections. IEEE Trans. on Pattern Analysis and Machine Intelligence, 8(2):222--233, March 1986.

....Radon transform, shift property. I. MOTION IN THE PROJECTION DOMAIN T HE SHIFT property of the Radon transform has found applications in many areas of image processing. For instance, in translational motion estimation from a video sequence [1] 2] and the related problem of image registration [3]. More importantly, projections acquired while the subject undergoes linear motion can be corrected using this property before a reconstruction of the image is attempted. The shift property of the Radon transform shows that translational motion in the image domain results in translational motion ....

S. Alliney and C. Morandi, "Digital image registration using projections, " IEEE Trans. Pattern Anal. Machine Intell., vol. PAMI-8, pp. 222--233, Mar. 1986.

....in determining its position may be poor [6] A way of sharpening the cross correlation peak is to whiten the input signals. A statistical approach to the whitening problem was proposed by Pratt [6] while a deterministic technique is the basis of the so called phasecorrelation technique [7] [10]. Such a method can be shown to be equivalent to adopting a whitening filter of FIR type, designed by means of the frequency sampling technique. The aim of this work is testing the robustness of the cross correlation and phase correlation techniques to small amounts of additive white noise. ....

.... notion in mind, the accuracy analysis will be performed considering the continuous crosscorrelation or phase correlation function, respectively obtained by interpolating (1) or (2) Assuming the input signals energies are normalized to the unit, by means of the Parseval theorem it can be seen [10] that the cross correlation and phase correlation algorithms are equivalent to the following procedures in the frequency domain: Cross correlation: Find 2 R which minimizes form L 2 Gamma1 X k=0 fi fi fiX 1 (k)X 2 (k) Gamma e Gammaj 2 k L T fi fi fi 2 (4) Phase correlation: ....

S. Alliney and C. Morandi, "Digital Image Registration Using Projections", IEEE Trans. Pattern Anal. Machine Intell., vol. 8, no. 2, pp. 222--233, 1986

No context found.

S. Alliney and C. Morandi, "Digital image registration using projections," IEEE Trans. Pattern Anal. Machine Intell., vol. 8, no. 2, pp. 222--233, Mar. 1986.

No context found.

S. Alliney, C. Morandi, \Digital Image Registration using Projections", IEEE Trans. on Pattern Analysis and Machine Intelligence, vol.8, n.2, pp. 222-233, 1986.

No context found.

S. Alliney and C. Morandi, "Digital image registration using projections," IEEE Trans. Pattern Anal. Machine Intell., vol. PAMI--8, no. 2, pp. 222--233, March 1986.

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