J. Ton and A. K. Jain, \Registering Landsat Images by Point Matching," IEEE Trans. Geosci. Remote Sensing, Vol. 27, No. 5, pp. 642-651, 1989.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....di#erent 14 acquisitions of the same finger have di#erent numbers of minutiae. A graph based representation [118, 155, 5] constructs a nearest neighbor graph from the minutiae patterns. The matching algorithm is based on inexact graph matching techniques. The point pattern based representation [11, 26, 96] considers the minutiae points as a two dimensional pattern of points. Correlation based techniques [61, 31] consider the gray level information in the fingerprint as features and match the global patterns of ridges and valleys to determine if the ridges align. The global representation of ....

J. Ton and A. K. Jain, "Registering Landsat Images by Point Matching," IEEE Trans. Geosci. Remote Sensing, Vol. 27, No. 5, pp. 642-651, 1989.

....considerable robustness towards noise, local as well as global landmark variation [Kitchen, 1985] Hence, these methods are better suited for our need. We utilise the basic point matching strategy [Ranade and Rosenfeld, 1980] but incorporate feature type information [Wang et al. 1983, Ton and Jain, 1989] and two way matching [Ton and Jain, 1989] The adapted, heuristic algorithm is presented in the remaining part of this section. Let A = fA 1 ; Delta Delta Delta ; Am g and B = fB 1 ; Delta Delta Delta ; Bn g represent the set of landmarks in both images. The basic idea of the algorithm is ....

....noise, local as well as global landmark variation [Kitchen, 1985] Hence, these methods are better suited for our need. We utilise the basic point matching strategy [Ranade and Rosenfeld, 1980] but incorporate feature type information [Wang et al. 1983, Ton and Jain, 1989] and two way matching [Ton and Jain, 1989]. The adapted, heuristic algorithm is presented in the remaining part of this section. Let A = fA 1 ; Delta Delta Delta ; Am g and B = fB 1 ; Delta Delta Delta ; Bn g represent the set of landmarks in both images. The basic idea of the algorithm is that if (A i ; B h ) is a true point pair, ....

Ton, J. and Jain, A. K. (1989). Registering landsat images by point matching. IEEE Transactions on Geoscience and Remote sensing, 27(5):642--651.

....landmark set and the landmarks in the sensed image(s) is to be established. Common diculty here is that there could be some landmarks having no corresponding counterparts in the other frame(s) Matching methods can be based on parameter clustering [30] graph matching [14] relaxation [26] [33] correlation like techniques [25] 2] 4] taking into account the landmark neighborhood or hybrid invariant approaches [34] 10] 35] 8] 6] This work has been supported by the grant No. 624178 3 of the Grant Agency of the Czech Ministry of Health. The authors thank to Dr. Igor ....

J. Ton and A. K. Jain. Registering Landsat images by point matching. IEEE Trans. Geoscience and Remote Sensing, 27(5):642-651, 1989.

....understanding of the energy function away from the optimal point leaves the way open for problems caused by local optima and slow convergence. Adhoc approaches are commonplace in the literature; for instance in probabilistic relaxation approaches, Ranganath and Chipman, 1992; Izumi et al. 1992; Ton and Jain, 1989) define support functions in an ad hoc fashion. In the structural domain (Horaud and Skordas, 1989) employ an empirical energy function for choosing between a number of candidate maximal cliques. It is also possible to impose syntactic constraints on the match such as demanding a oneto one ....

....The first step in discussing a relational graph matching technique is to establish a suitable formalism for describing the matching process. Here we describe the attributed relational graph (ARG) widely used in the literature (Barrow and Popplestone, 1971; Tang and Lee, 1992; Kittler et al. 1993; Ton and Jain, 1989). Our particular formulation is relatively simple; structural considerations are our primary concern here. A relational graph is represented by the triple G = V; E;X) and consists of a set of nodes V = fv 1 ; v 2 ; v n g which represent objects in a scene graph or model. The set E = fe 1 ; ....

[Article contains additional citation context not shown here]

Ton, J. and Jain, A. K. (1989). Registering landsat images by point matching. IEEE Transations on Geoscience and Remote Sensing, 27:642--651.

....images is trivial, and is done using a simple rule based system. Landmark types and position are then extracted using the architecture described above. Correspondence between landmarks in different images can be established in a number of ways. Current work is investigating relaxation schemes [7] and eigenvector analysis [5] of intra distances between landmark points. Empirical results are promising, allowing automated registration using a rigid body transformation (with 6 degrees of freedom) of corresponding points via a least squares technique [8] ....

J Ton and A K Jain. Registering landsat images by point matching. IEEE Transactions on Geoscience and Remonte sensing, 27(5):642--651, September 1989.

....and not the two way constraints required for many point matching problems. That is, there is a constraint that a point in one set can match to only one point in the other set, but there is no similar constraint for the points in the second set, i.e. there is no two way WTA (assignment) constraint. (Ton and Jain, 1989) attempt to impose such a two way constraint within the relaxation labeling framework but do not use other key techniques such as deterministic annealing, incorporated in algorithms described in this paper. Li, 1992) use a form of deterministic annealing (graduated non convexity) within the ....

Ton, J. and Jain, A. (1989). Registering Landsat images by point matching. IEEE Trans. Geo.

....sets can be matched using many techniques, including (i) point set matching [46] ii) graph matching [20] and (iii) sub graph isomorphism [1] These generic techniques are known to be intractable problems. Suboptimal solutions have been obtained using iterative procedures such as relaxation [44], simulated annealing, and genetic algorithms [3] The large computational requirement of matching is primarily due to the following three factors: i) a query fingerprint is usually of poor quality, ii) the fingerprint database is very large, and (iii) structural distortion of the fingerprint ....

....1992. 42] F. Stein and G. Medioni. Structural indexing: Efficient 2 D object recognition. IEEE Trans. on Pattern Analysis and Machine Intelligence, 14(12) 1198 1204, December 1992. 43] M. J. Swain and D. H. Ballard. Color indexing. International Journal of Computer Vision, 7(1) 11 32, 1991. [44] J. Ton and A. K. Jain. Registering Landsat images by point matching. IEEE Trans. on Geoscience and Remote Sensing, 27(5) 642 651, September 1989. 45] C. I. Watson. NIST special database 9: Mated Fingerprint Card Pairs. Advanced Systems Division, Image Recognition Group, National Institute for ....

J. Ton and A. K. Jain. Registering Landsat images by point matching. IEEE Trans. on Geoscience and Remote Sensing, 27(5):642--651, September 1989.

....first introduced by (Ranade and Rosenfeld, 1980) have also been widely applied to point matching. However these methods were originally developed as tools for classification and consequently in general do not impose the one to one matching constraints required in correspondence problems. (Ton and Jain, 1989) attempt to impose such a matching constraint within the relaxation labeling framework with limited success. Li, 1992) uses a form of deterministic annealing (graduated non convexity) within the relaxation labeling framework, but once again only with a one way matching constraint. Finally, ....

Ton, J. and Jain, A. (1989). Registering Landsat images by point matching. IEEE Trans. Geo. Rem. Sens., 27(5):642--651.

....O(l 3 m 2 ) complexity (where l and m are the number of links in the two graphs) though special instances are faster. The second approach employs nonlinear optimization methods (or heuristic approximations thereof) The most successful of these methods use some form of relaxation labeling [6, 7, 8, 9, 10, 11, 12, 13]. Relaxation labeling algorithms do not search the state space and generally have a much lower computational complexity (O(lm) or perhaps even lower see [10] than tree search methods. Other nonlinear optimization approaches are neural networks, 14, 15, 16, 17, 18, 19, 20, 21] linear ....

.... (or heuristic approximations thereof) The most successful of these methods use some form of relaxation labeling [6, 7, 8, 9, 10, 11, 12, 13] Relaxation labeling algorithms do not search the state space and generally have a much lower computational complexity (O(lm) or perhaps even lower see [10]) than tree search methods. Other nonlinear optimization approaches are neural networks, 14, 15, 16, 17, 18, 19, 20, 21] linear programming [22] symmetric polynomial transform [22] eigendecomposition [23] genetic algorithms [24] and Lagrangian relaxation [25] These techniques have so far ....

[Article contains additional citation context not shown here]

J. Ton and A. Jain, "Registering Landsat images by point matching", IEEE Trans. Geo. Rem. Sens., vol. 27, pp. 642--651, Sept. 1989.

....are faster. The second approach employs nonlinear optimization methods (or heuristic approximations thereof) The most successful of these methods use some form of relaxation labeling, like probabilistic relaxation (PR) Rosenfeld et al. 1976; Davis, 1979; Peleg, 1980; Hummel and Zucker, 1983; Ton and Jain, 1989; Li, 1992; Kittler et al. 1993) Probabilistic relaxation does not search the state space and generally has a much lower computational complexity (O(lm) or perhaps even lower see (Ton and Jain, 1989) than tree search methods. Other nonlinear optimization approaches are neural networks, ....

.... relaxation (PR) Rosenfeld et al. 1976; Davis, 1979; Peleg, 1980; Hummel and Zucker, 1983; Ton and Jain, 1989; Li, 1992; Kittler et al. 1993) Probabilistic relaxation does not search the state space and generally has a much lower computational complexity (O(lm) or perhaps even lower see (Ton and Jain, 1989)) than tree search methods. Other nonlinear optimization approaches are neural networks, Mjolsness et al. 1989; Mjolsness and Garrett, 1990; Simic, 1991; Yu and Tsai, 1992; Young et al. 1994; Chen and Lin, 1994) linear programming (Almohamad and Duffuaa, 1993) eigendecomposition (Umeyama, ....

[Article contains additional citation context not shown here]

Ton, J. and Jain, A. (1989). Registering Landsat images by point matching. IEEE Trans. Geo. Rem. Sens., 27(5):642--651.

....linear, affine) if D(A;B) cannot be decreased by applying to B a transformation from G. The main difference between various alignment approaches is in the distance function used: Huttenlocher et al. 14] use the Hausdorff distance, Sclaroff and Pentland [17] use strain energy , Ton and Jain [19] use support functions , and Horn [13] Besl and McKay [1] Gold et al. 9] and the statistical shape community [10] use a leastsquares type (Procrustes) distance. Other differences are the types of transformations allowed and whether point correspondences are set during the alignment process. We ....

J. Ton and A. K. Jain, "Registering Landsat images by point matching," IEEE Trans. Geosci. Remote Sensing, vol. 27,

No context found.

J. Ton and A. K. Jain, \Registering Landsat Images by Point Matching," IEEE Trans. Geosci. Remote Sensing, Vol. 27, No. 5, pp. 642-651, 1989.

....quality fingerprint contains between 60 and 80 minutiae, but different fingerprints have different number of minutiae. The variable sized minutiae based representation does not easily lend itself to indexing mechanisms. Further, typical graph based [9] 11] and point pattern based [1] 12] [13] approaches to match minutiae from two fingerprints need to align the unregistered minutiae patterns of different sizes which makes them computationally expensive. Correlation based techniques [14] 15] match the global patterns of ridges and valleys to determine if the ridges align. The global ....

J. Ton and A. K. Jain, "Registering landsat images by point matching," IEEE Trans. Geosci. Remote Sensing, vol. 27, pp. 642--651, May 1989.

....the correspondence problem. In remote sensing applications, point pattern matching is used for image registration. For the general problem of point pattern matching, where no a priori knowledge about the two sets of points is available, a number of algorithms have been described in the literature [17, 222, 205, 9, 221, 217]. Baird s O(n 2 ) algorithm, where n is the number of points in each of the two point sets, becomes more complex (O(n 3 ) when the number of points in the two sets are not the same. Vinod et al. 222] propose a neural network for point pattern matching after formulating it as a 0 1 integer ....

J. Ton and A. K. Jain. Registering Landsat images by point matching. IEEE Trans. on Geoscience and Remote Sensing, 27(5):642--651, September 1989.

....to solve the correspondence problem. In remote sensing applications, point pattern matching is used in image registration. For the general case of point pattern matching, where no a priori knowledge about the two sets of points is available, many algorithms have been described in the literature [1, 2, 3, 4, 5, 6]. Baird s O(n 2 ) algorithm, where n is the numbers of points in each of the two point sets, becomes more complex when the number of points are not same in the two sets. Vinod et al. 2] propose a neural network for point pattern matching after formulating the problem as a 0 1 integer ....

J. Ton and A. K. Jain, "Registering Landsat images by point matching," IEEE Trans. on Geoscience and Remote Sensing, vol. 27, pp. 642--651, September 1989.

....linear, ane) if D(A; B) cannot be further decreased by applying to B a transformation from G. The main di erence between various alignment approaches is in the distance function used: Huttenlocher et al. 8] use the Hausdor distance, Sclaro and Pentland [9] use strain energy , Ton and Jain [10] use support functions , and Horn [11] Besl and McKay [12] Gold et al. 13] and the statistical shape community [2] use a least squares type (Procrustes 1 ) distance. Other di erences are the types of transformations allowed 1 Procrustes was a villainous son of Poseidon in Greek mythology ....

J. Ton and A. K. Jain, \Registering Landsat images by point matching," IEEE Trans. Geosci. Remote Sensing, vol. 27, no. 5, pp. 642-651, 1989.

....Generally, an automatic fingerprint verification identification is achieved with point pattern matching (minutiae matching) instead of a pixel wise matching or a ridge pattern matching of fingerprint images. A number of point pattern matching algorithms have been proposed in the literature [23, 1, 21, 16]. Because a general point matching problem is essentially (a) input image (b) orientation field (c) fingerprint region (d) ridge map (e) thinned ridge map (f) extracted minutiae Figure 9: Results of our minutia extraction algorithm on a fingerprint image (512 Theta 512) captured with an inkless ....

....paths. The relaxation approach [16] iteratively adjusts the confidence level of each corresponding pair based on its consistency with other pairs until a certain criterion is satisfied. Although a number of modified versions of this algorithm have been proposed to reduce the matching complexity [23], these algorithms are inherently slow because of their iterative nature. The Hough transform based approach proposed by Stockman et al. 22] converts point pattern matching to a problem of detecting the highest peak in the Hough space of transformation parameters. It discretizes the ....

J. Ton and A. K. Jain, Registering Landsat Images by Point Matching, IEEE Transactions on Geoscience and Remote Sensing, Vol. 27, No. 5, pp. 642-651, 1989.

No context found.

J. Ton and A. K. Jain. Registering Landsat images by point matching. IEEE Trans. Geosci. Remote Sensing, 27(5):642--651, 1989.

....with an inkless scanner. matching of fingerprint images. Because a general point matching problem is essentially intractable, features associated with each point and their relative positions are widely used in the point pattern matching algorithms to reduce the exponential number of search paths [3, 1, 6, 9]. However, these algorithms are inherently slow and are unsuitable for an on line fingerprint verification system. In our system, an alignment based matching algorithm is implemented, which decomposes the minutia matching into two stages: i) alignment stage, where transformations such as ....

J. Ton and A. K. Jain, Registering Landsat Images by Point Matching, IEEE Transactions on Geoscience and Remote Sensing, Vol. 27, No. 5, pp. 642-651, 1989.

....algorithm is, instead, based on point pattern matching (minutiae matching) The reason for this choice is our need to design a robust, simple, and fast verification algorithm and to keep a small template size. A number of point pattern matching algorithms have been proposed in the literature [69, 1, 66, 55]. A general point matching problem is essentially intractable. Features associated with points and their spatial properties such as the relative distances between points are widely used in these algorithms to reduce the exponential number of search paths. The relaxation approach to point pattern ....

....to point pattern matching [55] iteratively adjusts the confidence level of each corresponding pair based on its consistency with other pairs until a certain criterion is satisfied. Although a number of modified versions of this algorithm have been proposed to reduce the matching complexity [69], these algorithms are inherently slow because of their iterative nature. The generalized Hough transform based approach to point pattern matching [6, 67] converts point pattern matching to a problem of detecting peaks in the Hough space of transformation parameters. It discretizes the parameter ....

J. Ton and A. K. Jain, Registering Landsat Images by Point Matching, IEEE Transactions on Geoscience and Remote Sensing, Vol. 27, No. 5, pp. 642-651, 1989.

....affine) if D(A; B) cannot be further decreased by applying to B a transformation from G. The main difference between various alignment approaches is in the distance function used: Huttenlocher et al. 8] use the Hausdorff distance, Sclaroff and Pentland [9] use strain energy , Ton and Jain [10] use support functions , and Horn [11] Besl and McKay [12] Gold et al. 13] and the statistical shape community [2] use a least squares type (Procrustes 1 ) distance. Other differences are the types of transformations allowed and whether point correspondences are established during the ....

J. Ton and A. K. Jain, "Registering Landsat images by point matching," IEEE Trans. Geosci. Remote Sensing, vol. 27, no. 5, pp. 642--651, 1989.

No context found.

Ton, J., Jain, A.K.: Registering landsat images by point matching. IEEE Transactions on Geoscience and Remote Sensing 27 (1989) 642--651

No context found.

J. Ton and A. K. Jain. Registering Landsat images by point matching. IEEE Transactions on Geoscience and Remote Sensing, 27:642--651, 1989. 28 (c) (d) (e)

No context found.

J. Ton and A. K. Jain. Registering Landsat images by point matching. IEEE Transactions on Geoscience and Remote Sensing, 27:642--651, 1989.

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