Skeleton Path Based Approach for Nonrigid 3D Shape Analysis and Retrieval

Size: px
Start display at page:

Download "Skeleton Path Based Approach for Nonrigid 3D Shape Analysis and Retrieval"

Transcription

1 Skeleton Path Based Approach for Nonrigid 3D Shape Analysis and Retrieval Chunyuan Li and A. Ben Hamza Concordia Institute for Information Systems Engineering Concordia University, Montréal, QC, Canada Abstract. In this paper, we propose a skeleton path based approach to analyze and retrieve nonrigid 3D shapes. The main idea is to match skeleton graphs by comparing the geodesic paths between skeleton endpoints. Our approach is motivated by the fact that the path feature is stable in the presence of articulation of components. The experimental results demonstrate the performance of our proposed method in terms of robustness to symmetry, discrimination against different graph structures, and high efficiency in nonrigid shape retrieval. Keywords: Skeleton Path, Nonrigid 3D Shapes, Retrieval. 1 Introduction With the increase in the number of scanned 3D objects, 3D shape analysis and retrieval is becoming popular in the fields of computer vision, computer graphics, and computer aided design. Previous efforts have been, however, mainly devoted to rigid 3D models, and thus how to efficiently and effectively analyze and compare nonrigid shapes is still a challenging problem. The curve skeleton, which integrates geometrical and topological features of the object, is an important shape descriptor. Shape similarity based on skeleton matching usually performs better than mesh surface or other shape descriptors in the presence of articulation of components, especially for non-rigid shape. As pointed out in [7], the curve-skeleton provides characteristics like part/component matching, registration and visualization, intuitiveness, and articulated transformation invariance. Nonrigid shape matching is one of the most challenging problems in contentbased 3D object retrieval. The aim of the SHREC 21 Shape Retrieval Contest of Non-rigid 3D Models is to evaluate and compare the effectiveness of different 3D retrieval methods [1]. For a 3D retrieval algorithm, the shape descriptor and the similarity discrimination are two key components. The global and local isometry-invariant descriptor proposed recently by Wu et al. [18] captured well the global and local information. Also, Agathos et al. [1] proposed a retrieval methodology based on a graph-based object representation. This method makes use of a meaningful new mesh segmentation and the Earth mover s distance (EMD) similarity measure. J.K. Aggarwal et al. (Eds.): IWCIA 211, LNCS 6636, pp , 211. c Springer-Verlag Berlin Heidelberg 211

2 Skeleton Path Based Approach for Shape Analysis and Retrieval 85 In recent years, several skeleton based shape analysis and retrieval methods have been proposed. Sundar et al. [16] encoded the geometric and topological information in the form of a skeleton graph and used graph matching techniques for skeleton matching and comparison. Cornea et al. [7] enhanced the framework by using a new skeletonization algorithm and an extension of the manyto-many matching algorithm. Au et al. [2] presented a fast and fully automatic correspondence algorithm that allows matching of a wide variety of shapes with semantically similar structures but with different geometric details. Specifically, they attempted to find a one-to-one semantic correspondence between two sets of feature nodes of curve skeletons. However, determining the similarity between two given shapes does not necessarily require finding an exact correspondence between their shape components. Our work is partly a 3D extension of the 2D path similarity skeleton graph matching approach proposed by Bai. et al [4]. Unlike [4], a major goal of our approach is to discover the symmetry instead of finding correspondences. Siddiqi et al. [15] introduced a medial surface based method and obtained state-of-the-art performance on McGill Articulated Shape Benchmark [15]. Partly motivated by Leonardo s Vitruvian Man, which describes the perfect human form in geometrical terms, we propose a skeleton path feature to represent each component of a non-rigid 3D shape, assuming that this feature descriptor is isometry-invariant, i.e. invariant to the object s variational representation, rotation, translation, scaling, and nonrigid bending. A skeleton path refers to the geodesic paths between two endpoints in the curve-skeleton, as shown in Fig. 1(a), where these shortest paths are represented as sequences of radii of the maximal balls at the corresponding skeleton points. We also benefit from the fact that the proportions of the curve skeleton length for different components are different and are almost constant. Although we do not explicitly consider the topological structure of the skeleton graphs, we do not, however, completely ignore this structure. It is worth pointing out that this topological structure is implicitly represented by the fact that overlapping parts of the geodesic paths are similar. Therefore, our approach is flexible enough to perform extremely well on nonrigid 3D shapes. The rest of the paper is organized as follows. Section 2 describes the proposed approach. The experimental results using the proposed algorithm are provided in Section 3. Finally, we conclude in Section 4. 2 Proposed Approach Our proposed method may be described as a two-phase approach: 1) Skeleton path acquisition, which includes curve-skeleton extraction, endpoint detection, and path construction. 2) Endpoints matching, which consists of finding an ordered sequence of end nodes and post-processing of this new sequence. For convenience and efficiency, we adopt the curve-skeleton extraction algorithm developed by Cornea et al. [8]. Nevertheless, our algorithm can be generally applied on curve-skeletons with satisfactory homotopic and centered properties [9].

3 86 C. Li and A. Ben Hamza (a) (b) Fig. 1. (a) Our skeleton path based algorithm describes a 3D cow model as many paths between end nodes. (b) Symmetric components that are discovered by our algorithm are represented by end nodes rendered with the same colors. Note that only part of the results is shown. The proposed curve-skeleton extraction algorithm works on a volumetric representation of a 3D object. It uses a generalized potential field generated by charges placed on the surface of the object. Given a 3D vector field, we use concepts from vector field visualization to identify two types of seed points that we will use to construct a curve-skeleton: critical points and high divergence points. Skeleton segments are discovered using a force-following algorithm on the underlying vector field, starting at each of the identified seed points. The force-following process evaluates the vector (force) value at the current point and moves in the direction of the vector with a small pre-defined step. At critical points, where the force vanishes, the initial directions are determined by evaluating the eigenvalues and eigenvectors of the Jacobian at the critical point. More details about computing the curve-skeleton can be found in [8]. 2.1 Skeleton Path We first describe the initial steps for building the skeleton graphs. The following definitions apply to continuous skeletons as well as to curve-skeletons of 3D models (composed of voxels).

4 Skeleton Path Based Approach for Shape Analysis and Retrieval 87 Definition 1. A skeleton point having only one adjacent point is an endpoint (the skeleton endpoint); a skeleton point having three or more adjacent points is a junction point. If a skeleton point is not an endpoint or a junction point, it is called a connection point. Definition 2. The endpoint in the skeleton graph is called an end node, and the junction point in the skeleton graph is called a junction node. Definition 3. The shortest path between a pair of end nodes on a skeleton graph is called a skeleton path. Based on the curve skeletons that are extracted using the method described above, we provide details on how we detect endpoints and construct the skeleton path. A. Skeleton Endpoint Detection Since the curve-skeleton consists of many segments that have two ends (i.e. segment endpoints), we detect the skeleton endpoints by considering the distance between these end nodes. We denote the set of all the N segment endpoints of an input skeleton by P = {p 1,p 2,..., p N }. For simplicity, let p k P denote the testing endpoint. Given a threshold ε RD,letQ = {q : p k q ε RD,q P, q p k } be the nearest neighbor endpoint set. We consider p k as a skeleton endpoint if the size of Q is and as a junction point if the size is larger than 2. The rest of the segment endpoints and all the other skeleton points are connected points (see Fig. 2). Fig. 2. Illustration of the critical skeleton points using our method on a 3D airplane (left). On the right, (a) endpoint, (b) connected points between two segments, (c) junction points, and (d) connected points in a segment. For better viewing, please see the original color pdf file. B. Skeleton Path Construction After endpoints are detected, we construct the skeleton path as the shortest path between two given nodes (see Fig. 1). Suppose there are N end nodes {v i } i=1,...,n in the skeleton graph G to be matched. Path Length Percentage. Let Γ =(γ ij )beann N path length matrix, where γ ij denotes the geodesic distance (shortest path length) from the i-th end

5 88 C. Li and A. Ben Hamza node v i to the j-th end node v j. To preserve scale-invariance, we normalize the matrix Γ by the overall curve skeleton length L skel. In this way, we obtain an N N length percentage matrix L =(l ij ), where l ij = γ ij /L skel. Path Radius Vector. Let p(v m, v n ) denote the skeleton path from v m to v n. We sample p(v m, v n )withm equidistant points, which are all skeleton points. Let R m,n (t) be the radius of the maximal ball at the skeleton point with index t in p(v m, v n ). We define a vector of the radii of the maximal balls at the M sample points on p(v m, v n ) as follows: R m,n = ( R m,n (t) ) t=1,2,...,m =(r 1,r 2,...,r M ). (1) The distance transform value for each point is equal to the radius of maximal inscribed ball. Suppose there are N voxels in the original 3D model, then to make the proposed method invariant to scale, we normalize R m,n (t) in the following way: DT (t) R m,n = N 1/N i=1 DT(s i), (2) where s i varies over all N voxels in the model [4]. Path Distance. The model dissimilarity between two skeleton paths is called a path distance. If R =(r i ) i=1,...,m and R =(r i ) i=1,...,m denote the vectors of radii of two model paths p(u, v)andp(u, v ) respectively, then the path distance is defined as ϕ ( p(u, v),p(u, v ) ) = M i=1 (r i r i )2 r i + r i + α (l l ) 2 l + l, (3) where l and l are the length percentages of p(u, v) andp(u, v ) respectively. The parameter α is a weight factor [4]. In order to make our representation scale invariant, the path lengths are normalized. We include the path length percentages in Eq. (3), since the percentage is not reflected in the sequences of radii (all paths are sequences of M radii). Thus, our path representation and the path distance are scale invariant. 2.2 Endpoints Matching Sorting endpoints by summing path length percentage. In a skeleton graph G with N end nodes {v i } i=1,...,n, each end node has the skeleton paths to all other end nodes in the graph. Let T (v i )= N j=1 l ij be the total path length percentages of the end node v i. Thus, given an end node v k, there is a corresponding length percentage T (v k ). We order all the end nodes in G following these percentages, by ranking an endpoint with a higher percentage at the top of the list. Therefore, we obtain an ordered end node sequence S = {u 1, u 2,...,u N }.

6 Skeleton Path Based Approach for Shape Analysis and Retrieval 89 Endpoints Distance. Let G and G be two graphs to be matched, and denote by S = {u i } i=1,...,k+1 and S = {u i } i=1,...,n+1 their respective ordered end node sequences, with K N. Similar to shape contexts [5], the matching cost c(u i, u j ) between u i and u j is based on the paths to all other end nodes in G and G that emanate from u i and u j, correspondingly. Then, we compute the path distances between the two sequences and obtain a matrix Φ = ( ϕ(u i, u j )) of the path distances computed using Eq. (3). As suggested by Bai et al. [4], we use the optimal subsequence bijection (OSB) to compute the dissimilarity value: c(u i, u j)=osb ( ϕ(u i, u j) ) (4) Global Matching. Using Eq. (4), we compute the total dissimilarity matrix C(G, G )= ( c(u i, u j )) 1 i K+1 between G and G using the Hungarian algorithm. For each end node v i in G, the Hungarian algorithm can find its corre- 1 j N+1 sponding end node v i in G.SinceG and G may have different numbers of end nodes, the total dissimilarity value should include a penalty for end nodes that did not find any partner. To achieve this, we simply add additional rows with a constant value κ so that C(G, G ) becomes a square matrix. This constant value κ is the average of all the other values in C(G, G ). The intuition behind using the Hungarian algorithm is that we want to have a globally consistent one-to-one assignment of all end nodes with possibly assigning some end nodes to κ, which represents a dummy node. This means that we seek a one-to-one correspondence of the end nodes in the skeleton graphs (with possibly skipping some nodes by assigning them to a dummy node). However, the Hungarian algorithm does not preserve the order of the matched sequences. This does not influence the final score, since we can change the order only for similar symmetric end nodes. This is also part of the reason why we can detect symmetric components. 3 Experimental Results In this section, we evaluate the performance of the proposed method in three parts: Symmetric components discovery, matching between different graph structures, and illustration of the recognition performance of our method on McGill Articulated Shape Database. 3.1 Symmetric Components Discovery Given an end node sequence S = {u 1, u 2,...,u N } of a skeleton G, weobtaina new sequence Ŝ by changing the order of certain end nodes in a set C G and compute the dissimilarity to the original. If the result C(S, Ŝ) islessthanagiven threshold ε C, we consider the components containing these end nodes in C to be symmetric. Obviously, for a given nonrigid 3D shape, it is possible that there are more than one such set C, whose size may be larger than 2. In other words, there exist many symmetric component groups and that each group might have more

7 9 C. Li and A. Ben Hamza Fig. 3. The curve skeleton and discovered symmetric components indicated by end nodes with the same color Table 1. The matrix of dissimilarity values between the skeleton graph with the corresponding end nodes exchanged and the original. The colored values are symmetric node pairs than 2 components. We now give a simple example illustrating our symmetry components discovery approach. Fig. 3 shows the curve skeleton and the results on a 3D cow model. The end nodes, displayed with the same color, indicate the symmetric components. As we observe, the left front leg and the right front leg are symmetric components, they are shown in yellow color. The back legs in green and horns are displayed in blue. We indexed the end nodes so that symmetric nodes are clear to be presented. The matrix with elements indicating the dissimilarity if the corresponding nodes are exchanged is shown in Table 1. The dissimilarities between two most symmetric end nodes are marked with colored numbers. Here, we choose the parameters ε C =.1,M =5,α= 1. Besides the symmetry discovery of the 3D cow model in Fig. 3, we tested the process on several other examples. In Fig. 4(a), we first discover the symmetry in both the hands and the legs of a dancer. And the head is not symmetric to any part of the human body. Obviously, our method finds the correct symmetric components even in large variability due to articulation. Secondly, Fig. 4(b) shows the result of an eight-leg octopus, which demonstrates that the proposed

8 Skeleton Path Based Approach for Shape Analysis and Retrieval 91 (a) (b) (c) (d) Fig. 4. (a) Symmetry discovery of a dancer; (b) Symmetry discovery of an octopus; (c) Symmetry discovery of a crab; (d) Symmetry discovery of a chair method is able to discover symmetric groups with numerous members. A crab with eight legs and two eyes in Fig. 4(c) illustrates that our method works correctly in the situation when there are many symmetric groups with different quantities of member. Finally, we show the result of a four-leg chair in Fig. 4(d). It demonstrates that the proposed method also performs well in the presence of a rigid shape, even when its skeleton graph is not a tree. 3.2 Matching Skeletons with Different Graph Structures For skeleton graphs with the same number of end nodes, they might have very different graph structure. Sometimes there are similar path radii vectors. But path length percentage will enhance the performance. For example, shapes like snake and spectacle have two endpoints in their curve-skeletons. Moreover, the skeleton graph of the snake is a tree, whereas the skeleton graph of the spectacle is not a tree. To evaluate the performance of our proposed algorithm on distinguishing the topological difference, we use a small database that contains four nonrigid 3D shapes: Two spectacles and two snakes as shown in Fig. 5. The parameter M for this database was set to M = 5. We also show the results with α = in Fig. 6(a) and α = 1 in Fig. 6(b). By observing the rankings, it is evident that both of them could discriminate the skeleton graphs with different structures. Although the shortest paths between end nodes of the two classes are similar, the proposed method is, however, able to distinguish the structural difference between a closed loop and a line better by considering the length percentages. No matter how shapes, e.g. spectacles or snakes, deform due to articulation, the length percentages are always almost constant. Moreover, as

9 92 C. Li and A. Ben Hamza Fig. 5. Top: Two spectacles and their curve-skeletons; Bottom: Two snakes and their curve-skeletons (a) (b) Fig. 6. Comparison between α =andα = 1 on a small database. The distance between query and the given shape is also displayed. a matter of fact, shapes in different classes have different length percentages, which lead to more effective discrimination. 3.3 Retrieval on McGill 3D Articulated Shape Database Based on the above experimental results, our algorithm is validated to be robust to symmetry and discriminative to different graph structures. We demonstrate it further on McGill Articulated Shape Database with 255 objects divided into ten categories, namely, Ants, Crabs, Spectacles, Hands, Humans, Octopuses, Pliers, Snakes, Spiders, and Teddy Bears. Sample models from this database are shown in Fig. 7.

10 Skeleton Path Based Approach for Shape Analysis and Retrieval 93 Fig. 7. Sample shapes from McGill Articulated Shape Database. Only two shapes for each of the 1 classes are shown. Skeleton Path based Methods. Retrieving shapes that are similar to a given query shape from a database involves shape matching. However, determining the similarity between two given shapes does not necessarily require finding an exact correspondence between their shape components. In this section, we extend the shape similarity measure discussed in Section 3 to shape retrieval, and we also propose five methods based on the skeleton path. In the sequel, we will use the following abbreviations: SH: We denote the method in section 3 as SH, since it uses the square matrix with penalty and Hungarian algorithm. SDP: We denote the method that uses the square matrix with penalty and dynamic programming algorithm as SDP. NSH: We denote the method that uses the matrix, which is not square and without penalty, and the Hungarian algorithm as NSH. EMS: We define the dissimilarity from the query to a shape in the dataset as the sum of minimum endpoint distance of the query to all endpoints of the latter, and denote it as EMS. PMS: We define the dissimilarity from the query to a shape in the dataset as the sum of minimum skeleton path distance of the query to all skeleton paths of the latter, and denote it as PMS. Here we use the parameters M =5andα = 5. In our comparative analysis, we have used the precision/recall curve to measure the retrieval performance. Ideally, this curve should be a horizontal line at unit precision. For each query shape, we use the first 77 returned shapes with descending similarity rankings (i.e., ascending Euclidean distance ranking), dividing them into 11 groups accordingly. The retrieval results of the 5 skeleton path based methods on the whole McGill Articulated Shape Database are shown in Fig. 8. Obviously, PMS provides a much better performance that the other methods because it fully exploits the original information that skeleton paths carry. By finding the minimum value of skeleton path distances, we might establish a potential corresponding relationship between paths. However, as a matter of fact, there is no veracious global path correspondence. As for EMS, the second best method, we assume that the endpoints with minimum distance are corresponding to each other, although it may fail to find the global endpoints correspondence. Furthermore, SH and SDP are almost neck and neck in terms of retrieval accuracy, and both are superior to NSH, which demonstrate that the penalty plays a key role in shape

11 94 C. Li and A. Ben Hamza Fig. 8. Precision-recall plot of the proposed skeleton path based methods discrimination. Our implementation was done in MATALB on a Intel Core 2 Duo with 2. GHz. To give an idea about the timing: constructing the skeleton path takes on average slightly less than 1 minute for a 3D model; most of this time is actually consumed by Dijkstra s algorithm for finding the shortest path between two end nodes. 4 Conclusions We proposed a skeleton path based technique that is able to detect symmetric components, discriminate different graph structure and retrieve nonrigid 3D shapes. We represented a nonrigid shape by a set of geodesic paths between skeleton endpoints. These paths were compared using sequence matching. By detecting symmetric components, our framework is shown to be consistent with human semanteme based on curve-skeleton. Also, we found that it is possible to discover multiple components in a symmetric group. In addition, the proposed approach could enhance the performance of distinguishing the topological difference. Finally, our skeleton path based approach is shown to be effective, efficient and easily understandable for articulated 3D shape retrieval. References 1. Agathos, A., Pratikakis, I., Papadakis, P., Perantonis, S., Azariadis, P., Sapidis, N.: Retrieval of 3D articulated objects using a graph-based representation. The Visual Computer 26, (21) 2. Au,O.K.-C.,Tai,C.-L.,Cohen-Or,D.,Zheng,Y.,Fu,H.:Electorsvotingforfast automatic shape correspondence. In: Proc. Eurographics, vol. 29 (21)

12 Skeleton Path Based Approach for Shape Analysis and Retrieval Au, O.K.-C., Tai, C.-L., Chu, H.-K., Cohen-Or, D., Lee, T.-Y.: Skeleton extraction by mesh contraction. ACM Transactions on Graphics 27 (28) 4. Bai, X., Latecki, L.J.: Path similarity skeleton graph matching. IEEE Trans. Pattern Analysis and Machine Intelligence 3, (28) 5. Belongie, S., Malik, J., Puzicha, J.: Shape matching and object recognition using shape contexts. IEEE Trans. Pattern Analysis and Machine Intelligence 24, (22) 6. Chen, D.-Y., Tian, X.-P., Shen, Y.-T., Ouhyoung, M.: On visual similarity based 3D model retrieval. Computer Graphics Forum 22, (23) 7. Cornea, N.D., Demirci, M.F., Silver, D., Shokoufandeh, A., Dickinson, S., Kantor, P.B.: 3D object retrieval using many-to-many matching of curve skeletons. In: Proc. Int. Conf. Shape Modeling and Applications, pp (25) 8. Cornea, N.D., Silver, D., Yuan, X., Balasubramanian, R.: Computing hierarchical curve-skeletons of 3D objects. The Visual Computer 21, (25) 9. Cornea, N.D., Silver, D., Min, P.: Curve-skeleton properties, applications, and algorithms. IEEE Trans. Visualization and Computer Graphics 13, (27) 1. Lian, Z., Godil, A., Fabry, T., Furuya, T., Hermans, J., Ohbuchi, R., Shu, C., Smeets, D., Suetens, P., Vandermeulen, D., Wuhrer, S.: SHREC 21 Track: Nonrigid 3D shape retrieval. In: Proc. Eurographics Workshop on 3D Object Retrieval, pp. 1 8 (21) 11. Kazhdan, M., Chazelle, B., Dobkin, D., Finkelstein, A., Funkhouser, T.: A reflective symmetry descriptor. In: Heyden, A., Sparr, G., Nielsen, M., Johansen, P. (eds.) ECCV 22. LNCS, vol. 2351, pp Springer, Heidelberg (22) 12. Shilane, P., Min, P., Kazhdan, M., Funkhouser, T.: The Princeton shape benchmark. In: Proc. Shape Modeling International, pp (24) 13. Shinagawa, Y., Kunii, T.L., Kergosien, Y.L.: Surface coding based on Morse theory. IEEE Computer Graphics and Applications 11, (1991) 14. Siddiqi, K., Shokoufandeh, A., Dickinson, S., Zucker, S.: Shock graphs and shape matching. International Journal of Computer Vision 35, (1999) 15. Siddiqi, K., Zhang, J., Macrini, D., Shokoufandeh, A., Bouix, S., Dickinson, S.: Retrieving articulated 3-D models using medial surfaces. Machine Vision and Applications 19, (28) 16. Sundar, H., Silver, D., Gagvani, N., Dickinson, S.: Skeleton based shape matching and retrieval. In: Proc. Shape Modeling International, pp (23) 17. Wang, Y.-S., Lee, T.-Y.: Curve skeleton extraction using iterative least squares optimization. IEEE Trans. Visualization and Computer Graphics 14, (28) 18. Wu, H.-Y., Zha, H., Luo, T., Wang, X.-L., Ma, S.: Global and local isometryinvariant descriptor for 3D shape comparison and partial matching. In: Proc. CVPR, pp (21) 19.

Salient Local 3D Features for 3D Shape Retrieval

Salient Local 3D Features for 3D Shape Retrieval Salient Local 3D Features for 3D Shape Retrieval Afzal Godil a, Asim Imdad Wagan b a National Institute of Standards and Technolog Gaithersburg, MD, USA b Dept of CSE, QUEST Universit Nawabshah, Pakistan

More information

3D Object Retrieval using Many-to-many Matching of Curve Skeletons

3D Object Retrieval using Many-to-many Matching of Curve Skeletons 3D Object Retrieval using Many-to-many Matching of Curve Skeletons Nicu D. Cornea 1 M. Fatih Demirci 2 Deborah Silver 1 Ali Shokoufandeh 2 Sven J. Dickinson 3 Paul B. Kantor 1 1 Rutgers University 2 Drexel

More information

Benchmarks, Performance Evaluation and Contests for 3D Shape Retrieval

Benchmarks, Performance Evaluation and Contests for 3D Shape Retrieval Benchmarks, Performance Evaluation and Contests for 3D Shape Retrieval Afzal Godil 1, Zhouhui Lian 1, Helin Dutagaci 1, Rui Fang 2, Vanamali T.P. 1, Chun Pan Cheung 1 1 National Institute of Standards

More information

Shape-Based Retrieval of Articulated 3D Models Using Spectral Embedding

Shape-Based Retrieval of Articulated 3D Models Using Spectral Embedding Shape-Based Retrieval of Articulated 3D Models Using Spectral Embedding Varun Jain and Hao Zhang GrUVi Lab, School of Computing Sciences, Simon Fraser University, Burnaby, British Columbia, Canada {vjain,

More information

Decomposing and Sketching 3D Objects by Curve Skeleton Processing

Decomposing and Sketching 3D Objects by Curve Skeleton Processing Decomposing and Sketching 3D Objects by Curve Skeleton Processing Luca Serino, Carlo Arcelli, and Gabriella Sanniti di Baja Institute of Cybernetics E. Caianiello, CNR, Naples, Italy {l.serino,c.arcelli,g.sannitidibaja}@cib.na.cnr.it

More information

Multi-view 3D retrieval using silhouette intersection and multi-scale contour representation

Multi-view 3D retrieval using silhouette intersection and multi-scale contour representation Multi-view 3D retrieval using silhouette intersection and multi-scale contour representation Thibault Napoléon Telecom Paris CNRS UMR 5141 75013 Paris, France napoleon@enst.fr Tomasz Adamek CDVP Dublin

More information

Measuring Cubeness of 3D Shapes

Measuring Cubeness of 3D Shapes Measuring Cubeness of 3D Shapes Carlos Martinez-Ortiz and Joviša Žunić Department of Computer Science, University of Exeter, Exeter EX4 4QF, U.K. {cm265,j.zunic}@ex.ac.uk Abstract. In this paper we introduce

More information

SHREC 10 Track: Non-rigid 3D Shape Retrieval

SHREC 10 Track: Non-rigid 3D Shape Retrieval Eurographics Workshop on 3D Object Retrieval (2010), pp. 1 8 I. Pratikakis, M. Spagnuolo, T. Theoharis, and R. Veltkamp (Editors) SHREC 10 Track: Non-rigid 3D Shape Retrieval Z. Lian 1,2, A. Godil 1, T.

More information

3D Shape Matching by Geodesic Eccentricity

3D Shape Matching by Geodesic Eccentricity Author manuscript, published in "Search in 3D 28, Anchorage, Alaska : United States (28)" DOI :.9/CVPRW.28.456332 3D Shape Matching by Geodesic Eccentricity Adrian Ion PRIP, Vienna University of Technology

More information

Shape Classification Using Regional Descriptors and Tangent Function

Shape Classification Using Regional Descriptors and Tangent Function Shape Classification Using Regional Descriptors and Tangent Function Meetal Kalantri meetalkalantri4@gmail.com Rahul Dhuture Amit Fulsunge Abstract In this paper three novel hybrid regional descriptor

More information

A Spectral Approach to Shape-Based Retrieval of Articulated 3D Models

A Spectral Approach to Shape-Based Retrieval of Articulated 3D Models A Spectral Approach to Shape-Based Retrieval of Articulated 3D Models Varun Jain, Hao Zhang GrUVi Lab, School of Computing Science, Simon Fraser University, Burnaby, BC, Canada Email: vjain,haoz@cs.sfu.ca

More information

Structure-oriented Networks of Shape Collections

Structure-oriented Networks of Shape Collections Structure-oriented Networks of Shape Collections Noa Fish 1 Oliver van Kaick 2 Amit Bermano 3 Daniel Cohen-Or 1 1 Tel Aviv University 2 Carleton University 3 Princeton University 1 pplementary material

More information

Retrieving Articulated 3-D Models Using Medial Surfaces and their Graph Spectra

Retrieving Articulated 3-D Models Using Medial Surfaces and their Graph Spectra Retrieving Articulated 3-D Models Using Medial Surfaces and their Graph Spectra Juan Zhang, Kaleem Siddiqi, Diego Macrini 2, Ali Shokoufandeh 3, and Sven Dickinson 2 McGill University, School of Computer

More information

Prototype-based Intraclass Pose Recognition of Partial 3D Scans

Prototype-based Intraclass Pose Recognition of Partial 3D Scans Prototype-based Intraclass Pose Recognition of Partial 3D Scans Jacob MONTIEL Hamid LAGA Masayuki NAKAJIMA Graduate School of Information Science and Engineering, Tokyo Institute of Technology Global Edge

More information

A New MPEG-7 Standard: Perceptual 3-D Shape Descriptor

A New MPEG-7 Standard: Perceptual 3-D Shape Descriptor A New MPEG-7 Standard: Perceptual 3-D Shape Descriptor Duck Hoon Kim 1, In Kyu Park 2, Il Dong Yun 3, and Sang Uk Lee 1 1 School of Electrical Engineering and Computer Science, Seoul National University,

More information

Fast trajectory matching using small binary images

Fast trajectory matching using small binary images Title Fast trajectory matching using small binary images Author(s) Zhuo, W; Schnieders, D; Wong, KKY Citation The 3rd International Conference on Multimedia Technology (ICMT 2013), Guangzhou, China, 29

More information

A Graph-based Shape Matching Scheme for 3D Articulated Objects

A Graph-based Shape Matching Scheme for 3D Articulated Objects A Graph-based Shape Matching Scheme for 3D Articulated Objects Min-Wen Chao, Chao-Hung Lin, I-Cheng Yeh, Chih-Chieh Chang, and Tong-Yee Lee National Cheng-Kung University, Taiwan Abstract In this paper,

More information

Contour-Based Large Scale Image Retrieval

Contour-Based Large Scale Image Retrieval Contour-Based Large Scale Image Retrieval Rong Zhou, and Liqing Zhang MOE-Microsoft Key Laboratory for Intelligent Computing and Intelligent Systems, Department of Computer Science and Engineering, Shanghai

More information

A METHOD FOR CONTENT-BASED SEARCHING OF 3D MODEL DATABASES

A METHOD FOR CONTENT-BASED SEARCHING OF 3D MODEL DATABASES A METHOD FOR CONTENT-BASED SEARCHING OF 3D MODEL DATABASES Jiale Wang *, Hongming Cai 2 and Yuanjun He * Department of Computer Science & Technology, Shanghai Jiaotong University, China Email: wjl8026@yahoo.com.cn

More information

Algorithms for 3D Isometric Shape Correspondence

Algorithms for 3D Isometric Shape Correspondence Algorithms for 3D Isometric Shape Correspondence Yusuf Sahillioğlu Computer Eng. Dept., Koç University, Istanbul, Turkey (PhD) Computer Eng. Dept., METU, Ankara, Turkey (Asst. Prof.) 2 / 53 Problem Definition

More information

Image retrieval based on bag of images

Image retrieval based on bag of images University of Wollongong Research Online Faculty of Informatics - Papers (Archive) Faculty of Engineering and Information Sciences 2009 Image retrieval based on bag of images Jun Zhang University of Wollongong

More information

3D Articulated Object Retrieval using a graph-based representation

3D Articulated Object Retrieval using a graph-based representation The Visual Computer manuscript No. (will be inserted by the editor) Alexander Agathos Ioannis Pratikakis Panagiotis Papadakis Stavros Perantonis Philip Azariadis Nickolas S. Sapidis 3D Articulated Object

More information

Similarity Shape Based on Skeleton Graph Matching

Similarity Shape Based on Skeleton Graph Matching Journal of Information Hiding and Multimedia Signal Processing c 2016 ISSN 2073-4212 Ubiquitous International Volume 7, Number 6, November 2016 Similarity Shape Based on Skeleton Graph Matching Truong-Giang

More information

On Skeletons Attached to Grey Scale Images. Institute for Studies in Theoretical Physics and Mathematics Tehran, Iran ABSTRACT

On Skeletons Attached to Grey Scale Images. Institute for Studies in Theoretical Physics and Mathematics Tehran, Iran ABSTRACT On Skeletons Attached to Grey Scale Images M. Karimi Behbahani, Arash Rafiey, 2 Mehrdad Shahshahani 3 Institute for Studies in Theoretical Physics and Mathematics Tehran, Iran ABSTRACT In [2], [3] and

More information

New Removal Operators for Surface Skeletonization

New Removal Operators for Surface Skeletonization New Removal Operators for Surface Skeletonization Carlo Arcelli, Gabriella Sanniti di Baja, and Luca Serino Institute of Cybernetics "E.Caianiello", CNR, Pozzuoli (Naples), Italy (c.arcelli, g.sannitidibaja,

More information

Skeleton-based Hierarchical Shape Segmentation

Skeleton-based Hierarchical Shape Segmentation Skeleton-based Hierarchical Shape Segmentation Dennie Reniers, Alexandru Telea Department of Mathematics and Computer Science Eindhoven University of Technology PO Box 513, 5600 MB Eindhoven, The Netherlands

More information

A New Method for Skeleton Pruning

A New Method for Skeleton Pruning A New Method for Skeleton Pruning Laura Alejandra Pinilla-Buitrago, José Fco. Martínez-Trinidad, and J.A. Carrasco-Ochoa Instituto Nacional de Astrofísica, Óptica y Electrónica Departamento de Ciencias

More information

A Comparative Study of Region Matching Based on Shape Descriptors for Coloring Hand-drawn Animation

A Comparative Study of Region Matching Based on Shape Descriptors for Coloring Hand-drawn Animation A Comparative Study of Region Matching Based on Shape Descriptors for Coloring Hand-drawn Animation Yoshihiro Kanamori University of Tsukuba Email: kanamori@cs.tsukuba.ac.jp Abstract The work of coloring

More information

Retrieving Articulated 3-D Models Using Medial Surfaces

Retrieving Articulated 3-D Models Using Medial Surfaces Machine Vision and Applications manuscript No. (will be inserted by the editor) Kaleem Siddiqi Juan Zhang Diego Macrini Ali Shokoufandeh Sylvain Bouix Sven Dickinson Retrieving Articulated 3-D Models Using

More information

Learning 3D Part Detection from Sparsely Labeled Data: Supplemental Material

Learning 3D Part Detection from Sparsely Labeled Data: Supplemental Material Learning 3D Part Detection from Sparsely Labeled Data: Supplemental Material Ameesh Makadia Google New York, NY 10011 makadia@google.com Mehmet Ersin Yumer Carnegie Mellon University Pittsburgh, PA 15213

More information

CS 468 Data-driven Shape Analysis. Shape Descriptors

CS 468 Data-driven Shape Analysis. Shape Descriptors CS 468 Data-driven Shape Analysis Shape Descriptors April 1, 2014 What Is A Shape Descriptor? Shapes Shape Descriptor F1=[f1, f2,.., fn] F2=[f1, f2,.., fn] F3=[f1, f2,.., fn] What Is A Shape Descriptor?

More information

Matching and Recognition in 3D. Based on slides by Tom Funkhouser and Misha Kazhdan

Matching and Recognition in 3D. Based on slides by Tom Funkhouser and Misha Kazhdan Matching and Recognition in 3D Based on slides by Tom Funkhouser and Misha Kazhdan From 2D to 3D: Some Things Easier No occlusion (but sometimes missing data instead) Segmenting objects often simpler From

More information

Accelerating Bag-of-Features SIFT Algorithm for 3D Model Retrieval

Accelerating Bag-of-Features SIFT Algorithm for 3D Model Retrieval Accelerating Bag-of-Features SIFT Algorithm for 3D Model Retrieval Ryutarou Ohbuchi, Takahiko Furuya 4-3-11 Takeda, Kofu-shi, Yamanashi-ken, 400-8511, Japan ohbuchi@yamanashi.ac.jp, snc49925@gmail.com

More information

Detecting Printed and Handwritten Partial Copies of Line Drawings Embedded in Complex Backgrounds

Detecting Printed and Handwritten Partial Copies of Line Drawings Embedded in Complex Backgrounds 9 1th International Conference on Document Analysis and Recognition Detecting Printed and Handwritten Partial Copies of Line Drawings Embedded in Complex Backgrounds Weihan Sun, Koichi Kise Graduate School

More information

Skeleton-based hierarchical shape segmentation

Skeleton-based hierarchical shape segmentation Skeleton-based hierarchical shape segmentation Reniers, D.; Telea, A.C. Published: 01/01/2006 Document Version Publisher s PDF, also known as Version of Record (includes final page, issue and volume numbers)

More information

Deformable Registration of Cortical Structures via Hybrid Volumetric and Surface Warping

Deformable Registration of Cortical Structures via Hybrid Volumetric and Surface Warping Deformable Registration of Cortical Structures via Hybrid Volumetric and Surface Warping Tianming Liu, Dinggang Shen, and Christos Davatzikos Section of Biomedical Image Analysis, Department of Radiology,

More information

Supplementary Material for Ensemble Diffusion for Retrieval

Supplementary Material for Ensemble Diffusion for Retrieval Supplementary Material for Ensemble Diffusion for Retrieval Song Bai 1, Zhichao Zhou 1, Jingdong Wang, Xiang Bai 1, Longin Jan Latecki 3, Qi Tian 4 1 Huazhong University of Science and Technology, Microsoft

More information

Improving 3D Shape Retrieval Methods based on Bag-of Feature Approach by using Local Codebooks

Improving 3D Shape Retrieval Methods based on Bag-of Feature Approach by using Local Codebooks Improving 3D Shape Retrieval Methods based on Bag-of Feature Approach by using Local Codebooks El Wardani Dadi 1,*, El Mostafa Daoudi 1 and Claude Tadonki 2 1 University Mohammed First, Faculty of Sciences,

More information

Human pose estimation using Active Shape Models

Human pose estimation using Active Shape Models Human pose estimation using Active Shape Models Changhyuk Jang and Keechul Jung Abstract Human pose estimation can be executed using Active Shape Models. The existing techniques for applying to human-body

More information

Reeb Graphs Through Local Binary Patterns

Reeb Graphs Through Local Binary Patterns Reeb Graphs Through Local Binary Patterns Ines Janusch and Walter G. Kropatsch Pattern Recognition and Image Processing Group Institute of Computer Graphics and Algorithms Vienna University of Technology,

More information

Translation Symmetry Detection: A Repetitive Pattern Analysis Approach

Translation Symmetry Detection: A Repetitive Pattern Analysis Approach 2013 IEEE Conference on Computer Vision and Pattern Recognition Workshops Translation Symmetry Detection: A Repetitive Pattern Analysis Approach Yunliang Cai and George Baciu GAMA Lab, Department of Computing

More information

Shape Matching and Recognition Using Group-Wised Points

Shape Matching and Recognition Using Group-Wised Points Shape Matching and Recognition Using Group-Wised Points Junwei Wang, Yu Zhou, Xiang Bai, and Wenyu Liu Department of Electronics and Information Engineering, Huazhong University of Science and Technology,

More information

Non-Experts Shape Modeling for Dummies

Non-Experts Shape Modeling for Dummies Non-Experts Shape Modeling for Dummies (Modeling with Interchangeable Parts) Alla Sheffer (joint work with Vladislav Kraevoy & Dan Julius) Motivation - Easy creation of 3D Content Currently 3D modeling

More information

DETC D FACE RECOGNITION UNDER ISOMETRIC EXPRESSION DEFORMATIONS

DETC D FACE RECOGNITION UNDER ISOMETRIC EXPRESSION DEFORMATIONS Proceedings of the ASME 2014 International Design Engineering Technical Conferences & Computers and Information in Engineering Conference IDETC/CIE 2014 August 17-20, 2014, Buffalo, New York, USA DETC2014-34449

More information

A New Approach for Shape Dissimilarity Retrieval Based on Curve Evolution and Ant Colony Optimization

A New Approach for Shape Dissimilarity Retrieval Based on Curve Evolution and Ant Colony Optimization Proc. Int. Conf. on Recent Trends in Information Processing & Computing, IPC A New Approach for Shape Dissimilarity Retrieval Based on Curve Evolution and Ant Colony Optimization Younes Saadi 1, Rathiah

More information

Nonrigid Surface Modelling. and Fast Recovery. Department of Computer Science and Engineering. Committee: Prof. Leo J. Jia and Prof. K. H.

Nonrigid Surface Modelling. and Fast Recovery. Department of Computer Science and Engineering. Committee: Prof. Leo J. Jia and Prof. K. H. Nonrigid Surface Modelling and Fast Recovery Zhu Jianke Supervisor: Prof. Michael R. Lyu Committee: Prof. Leo J. Jia and Prof. K. H. Wong Department of Computer Science and Engineering May 11, 2007 1 2

More information

Shape Recognition by Combining Contour and Skeleton into a Mid-Level Representation

Shape Recognition by Combining Contour and Skeleton into a Mid-Level Representation Shape Recognition by Combining Contour and Skeleton into a Mid-Level Representation Wei Shen 1, Xinggang Wang 2, Cong Yao 2, and Xiang Bai 2 1 School of Communication and Information Engineering, Shanghai

More information

GLOBAL SAMPLING OF IMAGE EDGES. Demetrios P. Gerogiannis, Christophoros Nikou, Aristidis Likas

GLOBAL SAMPLING OF IMAGE EDGES. Demetrios P. Gerogiannis, Christophoros Nikou, Aristidis Likas GLOBAL SAMPLING OF IMAGE EDGES Demetrios P. Gerogiannis, Christophoros Nikou, Aristidis Likas Department of Computer Science and Engineering, University of Ioannina, 45110 Ioannina, Greece {dgerogia,cnikou,arly}@cs.uoi.gr

More information

Graph Matching Iris Image Blocks with Local Binary Pattern

Graph Matching Iris Image Blocks with Local Binary Pattern Graph Matching Iris Image Blocs with Local Binary Pattern Zhenan Sun, Tieniu Tan, and Xianchao Qiu Center for Biometrics and Security Research, National Laboratory of Pattern Recognition, Institute of

More information

Top-k Keyword Search Over Graphs Based On Backward Search

Top-k Keyword Search Over Graphs Based On Backward Search Top-k Keyword Search Over Graphs Based On Backward Search Jia-Hui Zeng, Jiu-Ming Huang, Shu-Qiang Yang 1College of Computer National University of Defense Technology, Changsha, China 2College of Computer

More information

The Generalized Shape Distributions for Shape Matching and Analysis

The Generalized Shape Distributions for Shape Matching and Analysis The Generalized Shape Distributions for Shape Matching and Analysis Yi Liu 1, Hongbin Zha 1, Hong Qin 2 {liuyi, zha}@cis.pku.edu.cn, qin@cs.sunysb.edu National Laboratory on Machine Perception, Peking

More information

Human Shape Variation - An Efficient Implementation using Skeleton

Human Shape Variation - An Efficient Implementation using Skeleton Human Shape Variation - An Efficient Implementation using Skeleton Dhriti Sengupta 1, Merina Kundu 2, Jayati Ghosh Dastidar 3 Abstract It is at times important to detect human presence automatically in

More information

Weighted Neighborhood Sequences in Non-Standard Three-Dimensional Grids Parameter Optimization

Weighted Neighborhood Sequences in Non-Standard Three-Dimensional Grids Parameter Optimization Weighted Neighborhood Sequences in Non-Standard Three-Dimensional Grids Parameter Optimization Robin Strand and Benedek Nagy Centre for Image Analysis, Uppsala University, Box 337, SE-7505 Uppsala, Sweden

More information

Shape Descriptor using Polar Plot for Shape Recognition.

Shape Descriptor using Polar Plot for Shape Recognition. Shape Descriptor using Polar Plot for Shape Recognition. Brijesh Pillai ECE Graduate Student, Clemson University bpillai@clemson.edu Abstract : This paper presents my work on computing shape models that

More information

A Novel Extreme Point Selection Algorithm in SIFT

A Novel Extreme Point Selection Algorithm in SIFT A Novel Extreme Point Selection Algorithm in SIFT Ding Zuchun School of Electronic and Communication, South China University of Technolog Guangzhou, China zucding@gmail.com Abstract. This paper proposes

More information

SURFACE CONSTRUCTION USING TRICOLOR MARCHING CUBES

SURFACE CONSTRUCTION USING TRICOLOR MARCHING CUBES SURFACE CONSTRUCTION USING TRICOLOR MARCHING CUBES Shaojun Liu, Jia Li Oakland University Rochester, MI 4839, USA Email: sliu2@oakland.edu, li4@oakland.edu Xiaojun Jing Beijing University of Posts and

More information

Deformable 2D Shape Matching Based on Shape Contexts and Dynamic Programming

Deformable 2D Shape Matching Based on Shape Contexts and Dynamic Programming Deformable 2D Shape Matching Based on Shape Contexts and Dynamic Programming Iasonas Oikonomidis and Antonis A. Argyros Institute of Computer Science, Forth and Computer Science Department, University

More information

Reference Point Detection for Arch Type Fingerprints

Reference Point Detection for Arch Type Fingerprints Reference Point Detection for Arch Type Fingerprints H.K. Lam 1, Z. Hou 1, W.Y. Yau 1, T.P. Chen 1, J. Li 2, and K.Y. Sim 2 1 Computer Vision and Image Understanding Department Institute for Infocomm Research,

More information

DEFORMABLE MATCHING OF HAND SHAPES FOR USER VERIFICATION. Ani1 K. Jain and Nicolae Duta

DEFORMABLE MATCHING OF HAND SHAPES FOR USER VERIFICATION. Ani1 K. Jain and Nicolae Duta DEFORMABLE MATCHING OF HAND SHAPES FOR USER VERIFICATION Ani1 K. Jain and Nicolae Duta Department of Computer Science and Engineering Michigan State University, East Lansing, MI 48824-1026, USA E-mail:

More information

The correspondence problem. A classic problem. A classic problem. Deformation-Drive Shape Correspondence. Fundamental to geometry processing

The correspondence problem. A classic problem. A classic problem. Deformation-Drive Shape Correspondence. Fundamental to geometry processing The correspondence problem Deformation-Drive Shape Correspondence Hao (Richard) Zhang 1, Alla Sheffer 2, Daniel Cohen-Or 3, Qingnan Zhou 2, Oliver van Kaick 1, and Andrea Tagliasacchi 1 July 3, 2008 1

More information

Head Frontal-View Identification Using Extended LLE

Head Frontal-View Identification Using Extended LLE Head Frontal-View Identification Using Extended LLE Chao Wang Center for Spoken Language Understanding, Oregon Health and Science University Abstract Automatic head frontal-view identification is challenging

More information

Algorithm research of 3D point cloud registration based on iterative closest point 1

Algorithm research of 3D point cloud registration based on iterative closest point 1 Acta Technica 62, No. 3B/2017, 189 196 c 2017 Institute of Thermomechanics CAS, v.v.i. Algorithm research of 3D point cloud registration based on iterative closest point 1 Qian Gao 2, Yujian Wang 2,3,

More information

Object Recognition Robust under Translation, Rotation and Scaling in Application of Image Retrieval

Object Recognition Robust under Translation, Rotation and Scaling in Application of Image Retrieval Object Recognition Robust under Translation, Rotation and Scaling in Application of Image Retrieval Sanun Srisuky, Rerkchai Fooprateepsiri? and Sahatsawat Waraklang? yadvanced Machine Intelligence Research

More information

HIGH-LEVEL SEMANTIC FEATURE FOR 3D SHAPE BASED ON DEEP BELIEF NETWORKS.

HIGH-LEVEL SEMANTIC FEATURE FOR 3D SHAPE BASED ON DEEP BELIEF NETWORKS. HIGH-LEVEL SEMANTIC FEATURE FOR 3D SHAPE BASED ON DEEP BELIEF NETWORKS Zhenbao Liu 1, Shaoguang Cheng 1, Shuhui Bu 1,, Ke Li 2 1 Northwestern Polytechnical University, Youyi west road #127, Xi an 710072,

More information

A 3D Point Cloud Registration Algorithm based on Feature Points

A 3D Point Cloud Registration Algorithm based on Feature Points International Conference on Information Sciences, Machinery, Materials and Energy (ICISMME 2015) A 3D Point Cloud Registration Algorithm based on Feature Points Yi Ren 1, 2, a, Fucai Zhou 1, b 1 School

More information

Robot localization method based on visual features and their geometric relationship

Robot localization method based on visual features and their geometric relationship , pp.46-50 http://dx.doi.org/10.14257/astl.2015.85.11 Robot localization method based on visual features and their geometric relationship Sangyun Lee 1, Changkyung Eem 2, and Hyunki Hong 3 1 Department

More information

Shape Matching. Michael Kazhdan ( /657)

Shape Matching. Michael Kazhdan ( /657) Shape Matching Michael Kazhdan (601.457/657) Overview Intro General Approach Minimum SSD Descriptor Goal Given a database of 3D models, and given a query shape, find the database models that are most similar

More information

Query-Sensitive Similarity Measure for Content-Based Image Retrieval

Query-Sensitive Similarity Measure for Content-Based Image Retrieval Query-Sensitive Similarity Measure for Content-Based Image Retrieval Zhi-Hua Zhou Hong-Bin Dai National Laboratory for Novel Software Technology Nanjing University, Nanjing 2193, China {zhouzh, daihb}@lamda.nju.edu.cn

More information

Geometric Registration for Deformable Shapes 3.3 Advanced Global Matching

Geometric Registration for Deformable Shapes 3.3 Advanced Global Matching Geometric Registration for Deformable Shapes 3.3 Advanced Global Matching Correlated Correspondences [ASP*04] A Complete Registration System [HAW*08] In this session Advanced Global Matching Some practical

More information

Model-based segmentation and recognition from range data

Model-based segmentation and recognition from range data Model-based segmentation and recognition from range data Jan Boehm Institute for Photogrammetry Universität Stuttgart Germany Keywords: range image, segmentation, object recognition, CAD ABSTRACT This

More information

Robust 2D Shape Correspondence using Geodesic Shape Context

Robust 2D Shape Correspondence using Geodesic Shape Context Robust 2D Shape Correspondence using Geodesic Shape Context Varun Jain Richard (Hao) Zhang GrUVi Lab, School of Computing Science Simon Fraser University, Burnaby, BC, Canada E-mail: vjain, haoz@cs.sfu.ca

More information

Fully Automatic Methodology for Human Action Recognition Incorporating Dynamic Information

Fully Automatic Methodology for Human Action Recognition Incorporating Dynamic Information Fully Automatic Methodology for Human Action Recognition Incorporating Dynamic Information Ana González, Marcos Ortega Hortas, and Manuel G. Penedo University of A Coruña, VARPA group, A Coruña 15071,

More information

Non-rigid 3D Model Retrieval Using Set of Local Statistical Features

Non-rigid 3D Model Retrieval Using Set of Local Statistical Features Non-rigid 3D Model Retrieval Using Set of Local Statistical Features Yuki Ohkita, Yuya Ohishi, University of Yamanashi Kofu, Yamanashi, Japan {g09mk004, g11mk008}at yamanashi.ac.jp Takahiko Furuya Nisca

More information

Shape Similarity Measurement for Boundary Based Features

Shape Similarity Measurement for Boundary Based Features Shape Similarity Measurement for Boundary Based Features Nafiz Arica 1 and Fatos T. Yarman Vural 2 1 Department of Computer Engineering, Turkish Naval Academy 34942, Tuzla, Istanbul, Turkey narica@dho.edu.tr

More information

Registration of Dynamic Range Images

Registration of Dynamic Range Images Registration of Dynamic Range Images Tan-Chi Ho 1,2 Jung-Hong Chuang 1 Wen-Wei Lin 2 Song-Sun Lin 2 1 Department of Computer Science National Chiao-Tung University 2 Department of Applied Mathematics National

More information

Extracting Curve-Skeleton for Use in Shape Deformation

Extracting Curve-Skeleton for Use in Shape Deformation Journal of Computational Information Systems 9: 11 (2013) 4469 4478 Available at http://www.jofcis.com Extracting Curve-Skeleton for Use in Shape Deformation Long YANG 1, Zhiyi ZHANG 1,, Dongjian HE 2,

More information

Temperature Distribution Descriptor for Robust 3D Shape Retrieval

Temperature Distribution Descriptor for Robust 3D Shape Retrieval Temperature Distribution Descriptor for Robust 3D Shape Retrieval Yi Fang Purdue University West Lafayette, IN, USA fang4@purdue.edu Mengtian Sun Purdue University West Lafayette, IN, USA sun84@purdue.edu

More information

Robust Lip Contour Extraction using Separability of Multi-Dimensional Distributions

Robust Lip Contour Extraction using Separability of Multi-Dimensional Distributions Robust Lip Contour Extraction using Separability of Multi-Dimensional Distributions Tomokazu Wakasugi, Masahide Nishiura and Kazuhiro Fukui Corporate Research and Development Center, Toshiba Corporation

More information

Sketchable Histograms of Oriented Gradients for Object Detection

Sketchable Histograms of Oriented Gradients for Object Detection Sketchable Histograms of Oriented Gradients for Object Detection No Author Given No Institute Given Abstract. In this paper we investigate a new representation approach for visual object recognition. The

More information

Finding Median Point-Set Using Earth Mover s Distance

Finding Median Point-Set Using Earth Mover s Distance Proceedings of the Twenty-Eighth AAAI Conference on Artificial Intelligence Finding Median Point-Set Using Earth Mover s Distance Hu Ding and Jinhui Xu Computer Science and Engineering, State University

More information

Trajectory Compression under Network Constraints

Trajectory Compression under Network Constraints Trajectory Compression under Network Constraints Georgios Kellaris, Nikos Pelekis, and Yannis Theodoridis Department of Informatics, University of Piraeus, Greece {gkellar,npelekis,ytheod}@unipi.gr http://infolab.cs.unipi.gr

More information

Improving Image Segmentation Quality Via Graph Theory

Improving Image Segmentation Quality Via Graph Theory International Symposium on Computers & Informatics (ISCI 05) Improving Image Segmentation Quality Via Graph Theory Xiangxiang Li, Songhao Zhu School of Automatic, Nanjing University of Post and Telecommunications,

More information

Semi-Automatic Prediction of Landmarks on Human Models in Varying Poses

Semi-Automatic Prediction of Landmarks on Human Models in Varying Poses Semi-Automatic Prediction of Landmarks on Human Models in Varying Poses Stefanie Wuhrer Zouhour Ben Azouz Chang Shu National Research Council of Canada, Ottawa, Ontario, Canada {stefanie.wuhrer, zouhour.benazouz,

More information

Accelerating Pattern Matching or HowMuchCanYouSlide?

Accelerating Pattern Matching or HowMuchCanYouSlide? Accelerating Pattern Matching or HowMuchCanYouSlide? Ofir Pele and Michael Werman School of Computer Science and Engineering The Hebrew University of Jerusalem {ofirpele,werman}@cs.huji.ac.il Abstract.

More information

Similarity Image Retrieval System Using Hierarchical Classification

Similarity Image Retrieval System Using Hierarchical Classification Similarity Image Retrieval System Using Hierarchical Classification Experimental System on Mobile Internet with Cellular Phone Masahiro Tada 1, Toshikazu Kato 1, and Isao Shinohara 2 1 Department of Industrial

More information

Advanced Topics In Machine Learning Project Report : Low Dimensional Embedding of a Pose Collection Fabian Prada

Advanced Topics In Machine Learning Project Report : Low Dimensional Embedding of a Pose Collection Fabian Prada Advanced Topics In Machine Learning Project Report : Low Dimensional Embedding of a Pose Collection Fabian Prada 1 Introduction In this project we present an overview of (1) low dimensional embedding,

More information

A Graph Theoretic Approach to Image Database Retrieval

A Graph Theoretic Approach to Image Database Retrieval A Graph Theoretic Approach to Image Database Retrieval Selim Aksoy and Robert M. Haralick Intelligent Systems Laboratory Department of Electrical Engineering University of Washington, Seattle, WA 98195-2500

More information

Spatial Topology of Equitemporal Points on Signatures for Retrieval

Spatial Topology of Equitemporal Points on Signatures for Retrieval Spatial Topology of Equitemporal Points on Signatures for Retrieval D.S. Guru, H.N. Prakash, and T.N. Vikram Dept of Studies in Computer Science,University of Mysore, Mysore - 570 006, India dsg@compsci.uni-mysore.ac.in,

More information

ECCV Presented by: Boris Ivanovic and Yolanda Wang CS 331B - November 16, 2016

ECCV Presented by: Boris Ivanovic and Yolanda Wang CS 331B - November 16, 2016 ECCV 2016 Presented by: Boris Ivanovic and Yolanda Wang CS 331B - November 16, 2016 Fundamental Question What is a good vector representation of an object? Something that can be easily predicted from 2D

More information

Improving the Efficiency of Fast Using Semantic Similarity Algorithm

Improving the Efficiency of Fast Using Semantic Similarity Algorithm International Journal of Scientific and Research Publications, Volume 4, Issue 1, January 2014 1 Improving the Efficiency of Fast Using Semantic Similarity Algorithm D.KARTHIKA 1, S. DIVAKAR 2 Final year

More information

Learning the Three Factors of a Non-overlapping Multi-camera Network Topology

Learning the Three Factors of a Non-overlapping Multi-camera Network Topology Learning the Three Factors of a Non-overlapping Multi-camera Network Topology Xiaotang Chen, Kaiqi Huang, and Tieniu Tan National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy

More information

Dimension Reduction CS534

Dimension Reduction CS534 Dimension Reduction CS534 Why dimension reduction? High dimensionality large number of features E.g., documents represented by thousands of words, millions of bigrams Images represented by thousands of

More information

Improvement of SURF Feature Image Registration Algorithm Based on Cluster Analysis

Improvement of SURF Feature Image Registration Algorithm Based on Cluster Analysis Sensors & Transducers 2014 by IFSA Publishing, S. L. http://www.sensorsportal.com Improvement of SURF Feature Image Registration Algorithm Based on Cluster Analysis 1 Xulin LONG, 1,* Qiang CHEN, 2 Xiaoya

More information

Rotation Invariant Finger Vein Recognition *

Rotation Invariant Finger Vein Recognition * Rotation Invariant Finger Vein Recognition * Shaohua Pang, Yilong Yin **, Gongping Yang, and Yanan Li School of Computer Science and Technology, Shandong University, Jinan, China pangshaohua11271987@126.com,

More information

An explicit feature control approach in structural topology optimization

An explicit feature control approach in structural topology optimization th World Congress on Structural and Multidisciplinary Optimisation 07 th -2 th, June 205, Sydney Australia An explicit feature control approach in structural topology optimization Weisheng Zhang, Xu Guo

More information

Aligning Concave and Convex Shapes

Aligning Concave and Convex Shapes Aligning Concave and Convex Shapes Silke Jänichen and Petra Perner Institute of Computer Vision and applied Computer Sciences, IBaI, Körnerstr. 10, 04107 Leipzig pperner@ibai-institut.de, www.ibai-institut.de

More information

Image Registration with Local Rigidity Constraints

Image Registration with Local Rigidity Constraints Image Registration with Local Rigidity Constraints Jan Modersitzki Institute of Mathematics, University of Lübeck, Wallstraße 40, D-23560 Lübeck Email: modersitzki@math.uni-luebeck.de Abstract. Registration

More information

Intrinsic Mean for Semi-metrical Shape Retrieval Via Graph Cuts

Intrinsic Mean for Semi-metrical Shape Retrieval Via Graph Cuts Intrinsic Mean for Semi-metrical Shape Retrieval Via Graph Cuts Frank R. Schmidt 1,EnoTöppe 1,DanielCremers 1,andYuriBoykov 2 1 Department of Computer Science University of Bonn Römerstr. 164, 53117 Bonn,

More information

IMAGE RETRIEVAL USING VLAD WITH MULTIPLE FEATURES

IMAGE RETRIEVAL USING VLAD WITH MULTIPLE FEATURES IMAGE RETRIEVAL USING VLAD WITH MULTIPLE FEATURES Pin-Syuan Huang, Jing-Yi Tsai, Yu-Fang Wang, and Chun-Yi Tsai Department of Computer Science and Information Engineering, National Taitung University,

More information

Sketch-Based Articulated 3D Shape Retrieval

Sketch-Based Articulated 3D Shape Retrieval Feature Article Sketch-Based Articulated 3D Shape Retrieval Yusuf Sahilliog lu Middle East Technical University Metin Sezgin Koç University T hree-dimensional shape retrieval is an active area of research

More information