Q. Zheng and R. Chellappa. A computational vision approach to image registration. IEEE Trans. Image Processing, 2:311--326, 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....images from the union of smaller ones registered in their regions of overlap, is an old one. Mosaicking is closely related to image registration and rectification, and camera motion analysis. The technique of construction of mosaics by computer is also relatively old [6] Automated mosaicking [51] is often useful in its own right. Applications include station keeping [41] video coding [23] image stabilization [50] and visualization [42] Only recently have near real time [37] and globally consistent [36] mosaicking solutions emerged. 1.3.3 Prior Motion Estimation Work The literature ....

Q. Zheng and R. Chellappa. A computational vision approach to image registration. IEEE Transactions on Image Processing, 2(3):311-325, July 1993.

....fusion. The goal of image registration is to establish the spatial correspondence between two images. Several techniques have been developed for various types of data and applications [1] Existing techniques fall into two categories: intensitybased approaches [2, 3] and feature based techniques [4, 5]. In order to obtain a highly efficient and robust algorithm, pyramidal architectures are commonly utilized in the registration process [2, 3, 5, 6] In this paper, we present a Morphological Pyramid Image Registration (MPIR) algorithm that uses an intensity based differential method for ....

.... for various types of data and applications [1] Existing techniques fall into two categories: intensitybased approaches [2, 3] and feature based techniques [4, 5] In order to obtain a highly efficient and robust algorithm, pyramidal architectures are commonly utilized in the registration process [2, 3, 5, 6]. In this paper, we present a Morphological Pyramid Image Registration (MPIR) algorithm that uses an intensity based differential method for matching. This algorithm considers a model combining a 2 D affine transformation and an illumination change. The multiresolution images are represented by a ....

Q. Zheng and R. Chellappa, "A computational vision approach to image registration," IEEE Trans. Image Processing, vol.2, pp. 311-326, 1993.

....For example, spatial coordinates (landmarks) are well adapted to inter modal registration, where the purpose is to register two volumes measuring different properties of an object. However, the selection of landmarks is recognized to be a difficult problem [9, 10] whether done automatically [11] or manually [12] For many images, this is a serious drawback because registration accuracy can be no better than what is achieved by the initial selection of landmarks. For practical reasons, the number and precision of landmark locations is usually limited. Hence, spatial coordinates and ....

Q. Zheng and R. Chellappa, "A Computational Vision Approach to Image Registration," IEEE Transactions on Image Processing, vol. 2, no. 3, pp. 311--326, July 1993.

....is that motion can be compensated on demand, offering great flexibility by simply modifying some parameters of the compensation algorithm. 2.1. 2 Motion Estimation Our first prototype described in [9] was based on a multi resolution image registration technique developed by Zheng and Chellappa [20]. The technique matches a small set of feature points between two frames using a weighted correlation scheme and estimates four affine motion parameters using the computed feature displacements. The processing of each higher resolution level refines the estimates obtained from the registration of ....

....(SSD) of local patches around each feature. No estimation or refinement is done between different resolution levels, and the final displacements are used to estimate the motion parameters. In order to estimate the movement of the camera in a rigid environment we use the model described in [20]. To better understand the model, consider a camera mounted on a balloon. The camera is pointing down at a flat surface S, with its optical axis always perpendicular to S. In such a configuration, translations and rotations of the balloon will cause the image to translate and rotate ....

Q. Zheng and R. Chellappa. A computational vision approach to image registration. IEEE Transactions on Image Processing 2:311--326, 1993.

....automatic satellite image registration based on a multiresolution wavelet transform analysis has been implemented. A description of the mentioned algorithm and presentation of some satellite image registration results are the objectives of this paper. The technique is similar to that described in [16] but differs in some aspects. In the process of feature selection the algorithm uses feature points which are detected from the local modulus max ima of the wavelet transform. The correlation coefficient is used as a similarity measure and only the best pairwise fitting among all pairs of feature ....

....matches which are used to determine a warping function that gives the best registration of the LL subbands to the precision available in level L of the wavelet transform. To model the deformation between the images a 2 D affine transform with the parameters (s; Deltax; Deltay) is used [16]: X = T 1 (x; y) s[xcos( ysin( Deltax Y = T 2 (x; y) s[ Gammaxsin( ycos( Deltay; 11) where (x,y) and (X,Y) are corresponding points in the sensed and reference image, respectively. This model is commonly used in remote sensing applications and is a good approximation ....

Q. Zheng and R. A. Chellappa. computational vision approach to image registration. IEEE Transactions on Image Processing, 2(3):311--326, July 1993.

....body of research. In [5, 6] some correlation techniques are studied. The main focus is on how to reduce computational complexity, which has been the major factor preventing adoption of this approach. Recently, some new approaches were proposed in an attempt to reduce the computational complexity [7]. A number of alternative image registration approaches employ feature matching. Shape has been a commonly used feature for registration [8, 9] Region information can also be used to assist the registration [10] Some registration strategies using image features have also been studied [11] and ....

....which only affects the larger spatial frequencies of an image. This suggests a simple but effective approach for avoiding the problem which we call a coarse to fine (or multi scale) hierarchical approach. In fact, coarseto fine hierarchical approaches have been used by various researchers [7, 27, 28, 29, 30, 31, 32, 33]. It has been shown that hierarchical estimation techniques can often achieve efficient and robust performance. Here, we first estimate the optical flow using a coarse scale to avoid the potentially aliased high spatial frequencies. This first motion estimation is used as a correction to undo ....

Q. Zheng and R. Chellappa. A computational vision approach to image registration. IEEE Trans. Image Processing, 2(3):311--326, 1993.

.... geometries of the sensors, different spatial positions of the sensors, different temporal capture rates of the sensors and the inherent misalignment of the sensing elements [7] Registration is carried out either using optical flow techniques [1, 4, 5, 27, 32, 48] or feature matching techniques [44, 46, 60, 80]. These techniques align the images by exploiting the similarities (in graylevels or features) between the sensor images. The mismatch of image features in multisensor images reduces the similarities between the images and makes it difficult to establish the correspondence between the images. 13 ....

Q. Zheng and R. Chellappa. A computational vision approach to image registration. In Proceedings of the International Conference. on Pattern Recognition, pages 193-- 197. IEEE Computer Society Press, Los Alamitos, CA, 1992.

....scene, usually of unknown geometry. Mosaic based positioning deals with a two dimensional pose and assumes the scene geometry is flat, a fair assumption to make for floors. The idea of mosaicing, or building a large image from a collection of smaller ones, has been around for some time [5][6]. Applications for automated mosaicing have included station keeping [7] video coding [8] image stabilization [9] and visualization [10] Recently, real time mosaicing [11] and globally consistent mosaicing techniques [12] have emerged. However, there appears to be little literature on using ....

Q. Zheng and R. Chellappa, "A computational vision approach to image registration," IEEE Transactions on Image Processing, 2(3):311-325, July 1993.

....of combining three images, but in the final production, many more images were used, resulting in a high resolution full color composite showing most of the room (figure reproduced from [6] courtesy of IS T. TABLE I IMAGE COORDINATE TRANSFORMATIONS DISCUSSED IN THIS PAPER Zheng and Chellappa [3] considered the image registration problem using a subset of the affine model translation, rotation, and scale. Other researchers [4] 5] have assumed affine motion (six parameters) between frames. For the assumptions of static scene and no parallax, the affine model exactly describes rotation ....

Q. Zheng and R. Chellappa, "A computational vision approach to image registration," IEEE Trans. Image Processing, vol. 2, pp. 311--325, July 1993.

....shift (perhaps also disturbed by independent additive noise) This paper addresses the problem 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 ....

Q. Zheng and R. Chellappa. A computational vision approach to image registration. IEEE Trans. on Image Processing, 2(3):311--326, 1993.

....angle #. s a scaling parameter. T a translation vector. R a rotation matrix. s , T , and R the corrections for s, T, and R, respectively. 4 I. Introduction Image registration is an important technique for a great variety of applications such as aerial image analysis [1] 2] [3], stereo vision [4] 5] automated cartography [6] motion analysis[7] 8] and the recovery of the 3 D characteristics of a scene [9] There are two tasks which need to be handled during an image registration process. They are feature selection and correspondence establishment. Typically, ....

....of images must be preserved well. Therefore, their method is efficient but works well only on cases where the contour information is well preserved. On the other hand, the area based method usually adopts a window of points to determine a matched location using the correlation technique [1] [3] . The most commonly used measure is normalized cross correlation. This method is more robust than the feature based method in some situations. However, if the orientation difference between the two images is large, the value of cross correlation will be greatly influenced and the ....

[Article contains additional citation context not shown here]

Q. Zheng and R. Chellappa, "A computational vision approach to image registration," IEEE Trans. Image Processing, vol. 2, no. 3, pp. 311--326, July 1993.

No context found.

Q. Zheng and R. Chellappa. A computational vision approach to image registration. IEEE Trans. Image Processing, 2:311--326, 1993.

.... I to I can be written as I = T I I is the image obtained by mapping the points in I to the coordinate system of I and T is a 2 D to 2 D transformation of the form T : R Gamma R A number of choices are available for the 2 D transformation between two images, as discussed in [16, 18]. In practice, the form of T is chosen based on experience and knowledge of sensor geometries. A general transformation of a 2 D point set I into the coordinate system of another, I, can be written as I = T (a 1 ;a 2 ; a n ) I; where a 1 ; a 2 ; a n are the n parameters of the ....

....and then estimate the transformation parameters using (15) Almost invariably, there will be some false matches that will lead to errors in the estimates. A number of methods are available to detect and prune out such outliers [2] We use the iterative refinement approach developed in [18]. In this approach, transformation parameters are first estimated using the available candidate point correspondences. The computed transformation parameters are used to project points in the first image onto the second. A match (p 1 ; p 2 ) is considered to be correct if the projection of p 1 ....

Q. Zheng and R. Chellappa, "A computational vision approach to image registration," IEEE Transactions on Image Processing, Vol. 2, pp. 311-- 326, July 1993. 35

....significant structures, registering such images is a challenging and difficult problem. Tradi 7 tional solutions to this problem are unreliable when the rotation and scale change between the two frames is significant. A computational vision approach to solve this registration is proposed in [8] [19] using the feature detection algorithm described earlier. A small number of feature points are located in the two images from the sequence using the algorithm described in the previous section. Typically, 20 40 features are identified per image. The figure shows these feature locations (marked ....

....initial estimate of translation and scaling is obtained by pair wise matching of the detection feature points. Subsequently, a hierarchical correlation matching is performed to obtain an accurate camera motion estimate. For more details about this registration scheme the reader is referred to [8] [19]. It has been tested on many different data sets including stereo image pairs and satellite image data. This application illustrates the robustness of this feature detection method in identifying a consistent set of features irrespective of significant amounts of rotation, scaling and perspective ....

Q. Zheng and R. Chellappa, "A computational vision approach to image registration," IEEE Trans. Image Processing, Vol. 2, No. 3, pp. 311--326, July 1993.

....site model refinement, or verification purposes. Careful positioning of newly acquired images is therefore of paramount importance in a model supported exploitation paradigm. Relevant registration techniques have been described in works such as [Beveridge and Riseman, 1992; Collins et al. 1993; Zheng and Chellappa, 1993] The emphasis here is on automatic methods. The semiautomatic camera resection algorithm automatically extracts corners corresponding to the intersections of lines. These are chosen as possible image locations of 3D control points. The user can select the correct control point correspondences. ....

Q. Zheng and R. Chellappa. A computational vision approach to image registration. IEEE Trans. on Image Processing, 2:311--326, 1993. 27 28 (a) Initial projection of control points on the new image (b) Corners extracted from the new image

No context found.

Zheng Q., Chellapa R., "A Computational Vision Approach to Image Registration," IEEE Trans. on Image Processing, Vol.2, No.3, pp.311-326, 1993.

No context found.

Q. Zheng and R. A. Chellappa. computational vision approach to image registration. IEEE Transactions on Image Processing, 2(3):311--326, July

No context found.

Q. Zheng and R. Chellappa, "A Computational Vision Approach to Image Registration," IEEE IP, 2(3): 311-326, 1993.

No context found.

Q. Zheng and R. Chellappa, "A computational vision approach to image registration," IEEE Trans. Image Processing, vol. 2, no. 3, July 1993.

No context found.

Q. Zheng., R. Chellappa. "A computational vision approach to image registration", IEEE Transactions on Image Processing, 2(3), pp. 311-326, 1993.

No context found.

Q. Zheng and R. A. Chellappa. computational vision approach to image registration. IEEE Transactions on Image Processing, 2(3):311--326, July 1993.

No context found.

Q. Zheng and R. Chellapa, "A computational vision approach to image registration," IEEE Transactions on Image Processing, vol. 2(3), pp. 311--325, July 1993.

No context found.

Zheng, Q.; Chellappa, R. A computational vision approach to image registration. IEEE Trans. Image Processing 2:311-326; 1993.

No context found.

Q. Zheng, R. Chellappa, "A computational vision approach to image registration", IEEE Transactions on Image Processing, 1993, 2(3):311-326.

No context found.

Q. Zheng, R. Chellappa, "A computational vision approach to image registration", IEEE Transactions on Image Processing, 1993, 2(3):311-326.

First 50 documents

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