M.P. Dubuisson and A.K. Jain. A modified Hausdor# distance for object matching. In ICPR94, pages A:566--568, Jerusalem, Israel, 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....on dissimilarity relations between ob ects. Such representations are useful when features are di#cult to obtain or when they have little discriminative power. Such situations are encountered in practice, especially when shapes, blobs, or some particular image characteristics have to be recognized [6,8]. The use of dissimilarities is, therefore, dictated by the application or data specification. For an understanding of dissimilarity data, techniques of multidimensional scaling (MDS) 1,10] can be used. MDS refers to a group of methods mainly used for visualizing the structure in ....

....the classification experiments conducted, presents some 2D pro ection maps and discusses the results. Conclusions are summarized in section 6. 2 Linear Projection of the Dissimilarity Data Non metric distances may arise when shapes or ob ects in images are compared e.g. by template matching [8,6]. For pro ection purposes, the symmetry condition is necessary, but for any symmetric distance matrix, an Euclidean space is not large enough for a distance preserving linear mapping onto the specified dimensionality. It is, however, always possible [4] for a pseudo Euclidean space. The ....

[Article contains additional citation context not shown here]

Dubuisson M. P. and Jain A. K. Modified Hausdor# distance for object matching . In 12th Int. Conf. on Pattern Recognition, volume 1, pag9 566--568, 1994. 488, 489, 492

....digits 3 and 8 [10] The digits are represented as 128128 binary images. Since no natural features arise from the application, constructing dissimilarities is an interesting possibility to deal with such a recognition problem. Three dissimilarity measures are considered: Hamming, modified Hausdor# [6] and blurred , resulting in the representations: DH , DMH and DB correspondingly. The Hamming distance counts the number of pixels which disagree. The modified Hausdor# distance is found useful for template matching purposes [6] It measures the di#erence between two sets (here two contours) A ....

....dissimilarity measures are considered: Hamming, modified Hausdor# [6] and blurred , resulting in the representations: DH , DMH and DB correspondingly. The Hamming distance counts the number of pixels which disagree. The modified Hausdor# distance is found useful for template matching purposes [6]. It measures the di#erence between two sets (here two contours) A = a 1 , a g and B = b 1 , b h and is defined as DMH (A, B) max(hM (A, B) hM (B, A) where hM (A, B) 1 g # a#A min b#B a b . To find 0 0.2 0.4 SPEARMAN COEF. Blurred Mod.Hausd. 0 0.2 ....

Dubuisson M. P. and Jain A. K. Modified hausdor# distance for object matching. In 12th Int. Conf. on Pattern Recognition, volume 1, pages 566--568, 1994.

....are given as binary images with the resolution of 128 128 pixels. Since no natural features arise from the application, constructing dissimilarities is an interesting possibility to deal with such a recognition problem. Three dissimilarity measures are considered: Hamming, modified Hausdor# [6] and blurred , resulting in three representations: DH , DMH and DB correspondingly. The Hamming distance simply counts the number of pixels which disagree, i.e. have different binary values. The Hausdor# distance is often used for image comparison. Here, we will use a variant, namely the ....

....The Hamming distance simply counts the number of pixels which disagree, i.e. have different binary values. The Hausdor# distance is often used for image comparison. Here, we will use a variant, namely the modified Hausdor# distance since it is found more useful for template matching purposes [6]. We apply it on the contours of digits. The modified Hausdor# distance measures the di#erence between two sets A = a 1 , a g and B = b 1 , b h (here two contours) and is defined as DMH (A, B) max(hM (A, B) hM (B, A) where hM (A, B) 1 g # a#A min b#B a ....

Dubuisson M. P. and Jain A. K. Modified hausdor# distance for object matching. In 12th Int. Conf. on Pattern Recognition, volume 1, pages 566--568, 1994.

....quality is found upon division of the smallest side of each triangle by its largest side. If the triangle is equilateral this expression is equal to 1; The accuracy is expressed by the modified Hausdor# distance that represents the mean distance of the generated mesh to a reference shape [4]: H ave (S 1 , S 2 ) 1 N # p#S1 e(p, S 2 ) 7) where e is the minimum distance between a point and a surface, and S 1 and S 2 are two surfaces. Using these measures, the following experiment is conducted. Two volumes, one containing a greyscale image of a sphere and the other containing a ....

Dubuisson, M., and Jain, A. A modified Hausdor# distance for object matching. In Proceedings, 12th IAPR International Conference on Pattern Recognition, Conference A (Jerusalem, Israel, October 9--13, 1994) (Oct. 1994), IEEE Computer Society Press, Los Alamitos, CA, 1994, 566-568, pp. 566-- 568.

No context found.

M.P. Dubuisson and A.K. Jain. A modified Hausdor# distance for object matching. In ICPR94, pages A:566--568, Jerusalem, Israel, 1994.

....localization facemodel Fig. 1. Face finding procedure An e#cient yet powerful method to calculate the similarity of two binary images is the Hausdor# distance [7] a metric between two point sets. We use a slightly adapted measure, called the (directed) modified Hausdor# distance (MHD) [2] to calculate the similarity between the image and the model. Given the two point sets and B and some underlying norm on the points, the MHD is defined as hmod (A, B) 1 A a#A min b#B #a b# . 1) With the two dimensional point set representing the image and T p (B) ....

M.P. Dubuisson and A.K. Jain. A modified Hausdor# distance for object matching. In ICPR94, pages A:566--568, Jerusalem, Israel, 1994.

....the directed Hausdor# distance from set A to B with some underlying norm on the points of A and B. For image processing applications it has proven useful to apply a slightly di#erent measure, the (directed) modified Hausdor# distance (MHD) which was introduced by Dubuisson et al. [1]. It is defined as hmod (A, B) 1 A # a#A min b#B #a b# . 3) By taking the average of the single point distances, this version decreases the impact of outliers making it more suitable for pattern recognition purposes. 2.2 Model Based detection Rucklidge [4] describes a method ....

M.P. Dubuisson and A.K. Jain. A modified Hausdor# distance for object matching. In ICPR94, pages A:566--568, Jerusalem, Israel, 1994.

No context found.

M.-P. Dubuisson and A. K. Jain. A modified hausdor# distance for object matching. In 12th International Conference on Pattern Recognition, pages 566--568, Jerusalem, Israel, 1994.

No context found.

M.-P. Dubuisson and A. K. Jain, \A modied hausdor distance for object matching," Proceedings of the 12th International Conference on Pattern Recognition, Jerusalem, Israel 1994.

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