3DOR 11

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/310

Browse

Browsing 3DOR 11 by Title

Now showing 1 - 18 of 18
  • Thumbnail Image
    Item
    Bag of Words and Local Spectral Descriptor for 3D Partial Shape Retrieval
    (The Eurographics Association, 2011) Lavoué, Guillaume; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    This paper presents a 3D shape retrieval algorithm based on the Bag of Words (BoW) paradigm. For a given 3D shape, the proposed approach considers a set of feature points uniformly sampled on the surface and associated with local Fourier descriptors; this descriptor is computed in the neighborhood of each feature point by projecting the geometry onto the eigenvectors of the Laplace-Beltrami operator, it is highly discriminative, robust to connectivity and geometry changes and also fast to compute. In a preliminary step, a visual dictionary is built by clustering a large set of feature descriptors, then each 3D shape is described by an histogram of occurrences of these visual words. The performances of our approach have been compared against very recent state-of-theart methods on several different datasets. For global shape retrieval our approach is comparable to these recent works, however it clearly outperforms them in the case of partial shape retrieval.
  • Thumbnail Image
    Item
    ConTopo: Non-Rigid 3D Object Retrieval using Topological Information guided by Conformal Factors
    (The Eurographics Association, 2011) Sfikas, Konstantinos; Pratikakis, Ioannis; Theoharis, Theoharis; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    Combining the properties of conformal geometry and graph-based topological information for 3D object retrieval, a non-rigid 3D object descriptor is proposed, which is both robust and efficient in terms of retrieval accuracy and computation speed. In previous works, graph-based methods for non-rigid 3D object retrieval, have shown high discriminative power and robustness, while geometry-based methods, have proven to be tolerant to noise and pose. In this work, we present a 3D object descriptor that combines the above advantages.
  • Thumbnail Image
    Item
    Evaluation of 3D Interest Point Detection Techniques
    (The Eurographics Association, 2011) Dutagaci, Helin; Cheung, Chun Pan; Godil, Afzal; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    In this paper, we compare the results of five 3D interest point detection techniques to the interest points marked by human subjects. This comparison is used to quantitatively evaluate the interest point detection algorithms. We asked human subjects to look at a number of 3D models, and mark interest points on the models via a web-based interface. We propose a voting-based method to construct ground truth out of humans' selections of interest points. Evaluation measures, namely False Positive and False Negative Errors, are then defined based on the geodesic distance between the interest points detected by a particular algorithm and the human-generated ground truth.
  • Thumbnail Image
    Item
    Feature Template based 3D Model Retrieval
    (The Eurographics Association, 2011) Pan, Xiang; Chen, QiHua; Liu, Zhi; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    3D model retrieval has attracted more and more research interests. Lots of shape descriptors have been proposed till now. But during the process of constructing these shape descriptors, feature correlation among models is not considered. In this paper, we propose a simple but very effective method in improving retrieving accuracy by employing correlative information, namely feature template. Feature template is designed to remove such small variation while remaining discriminative features by performing meaning operation of feature vectors. As a result, it makes the feature vector be more robust for better retrieving accuracy. In addition, the feature template can be regarded as a post-processing of existing shape descriptors. Therefore, the proposed method can be used to improve retrieving accuracy for any shape descriptors in the form of feature vector. In experiments, we test the proposed method for several shape descriptors by using a public 3D model database. Comparing with original shape descriptors, our method can greatly improve the retrieval accuracy.
  • Thumbnail Image
    Item
    A Framework For 3D Object Retrieval Algorithm Analysis
    (The Eurographics Association, 2011) Lourenço, Tiago; Ferreira, Alfredo; Fonseca, Manuel J.; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    The increasing number of three-dimensional objects available on digital format triggered a great interest in research in this domain. Finding efficient methods of analysis, comparison and retrieval of 3D models has become an important task. However, despite the existence of some benchmarks with collections of 3D models, annual contests with specific tracks to compare techniques, and even a framework online (MMW.com) which allows to compare the performance of descriptors, there is no integrated system that provides, in a centralized manner, the necessary tools to study and compare the various techniques associated with 3D object retrieval. In this article, we present a modular and scalable web-based system that allows the addition of new components, like shape descriptors or segmentation algorithms, with minor effort by researchers who developed them.
  • Thumbnail Image
    Item
    Generative Object Definition and Semantic Recognition
    (The Eurographics Association, 2011) Ullrich, Torsten; Fellner, Dieter W.; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    ''What is the difference between a cup and a door?'' These kinds of questions have to be answered in the context of digital libraries. This semantic information, which describes an object on a high, abstract level, is needed in order to provide digital library services such as indexing, markup and retrieval. In this paper we present a new approach to encode and to extract such semantic information. We use generative modeling techniques to describe a class of objects: each class is represented by one algorithm; and each object is one set of high-level parameters, which reproduces the object if passed to the algorithm. Furthermore, the algorithm is annotated with semantic information, i.e. a human-readable description of the object class it represents. We use such an object description to recognize objects in real-world data e.g. laser scans. Using an algorithmic object description, we are able to identify 3D subparts, which can be described and generated by the algorithm. Furthermore, we can determine the needed input parameters. In this way, we can classify objects, recognize them semantically and we can determine their parameters (cup's height, radius, etc.).
  • Thumbnail Image
    Item
    A GPU Based High-efficient and Accurate Optimal Pose Alignment Approach of 3D Objects
    (The Eurographics Association, 2011) Zhang, Qian; Jia, Jinyuan; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    In this paper we present a new method for alignment of 3D objects. This approach is based on the exhaustive optimization search in the 3D space using GPU based genetic algorithm. The descriptor of 3D object used as the objective function to be optimized is a newly developed pose-variant similarity measure, which is obtained directly from the voxelized model's geometry and could be entirely implemented on the GPU. In order to reduce the traditional optimal algorithms' large processing time, we exploit the GPU's highly parallel architecture and transport our approach from CPU to GPU. Experimental results show that the proposed method is superior to existing normalization techniques such as PCA and provides a high degree of precision to align 3D objects.
  • Thumbnail Image
    Item
    Heat Diffusion Approach for Feature-based Body Scans Analysis
    (The Eurographics Association, 2011) Lovato, Christian; Zancanaro, C.; Castellani, Umberto; Giachetti, Andrea; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    In this paper we propose the use of spectral shape analysis techniques for selection and classification of anthro- pometric points extracted from human-body scans. Few feature points are detected by exploiting the capability of heat diffusion process in capturing the extremities of surface protrusions which are often related to anthropometric landmarks. Then, a heat kernel signature is computed for each feature point which are associated to its seman- tic group by employing a learning-by-example procedure exploiting manual point labeling provided by an expert anthropometrist. Detected points are not clearly the same precise anatomical locations used in standard anthro- pometric procedures, but their matching can be useful for different applications like automatic model registration or simple body type evaluation. Experimental tests carried out on several subjects with different anthropometric characteristics show encouraging results demonstrating the potential usefulness of the approach as well as the necessity of further investigation on point description and matching.
  • Thumbnail Image
    Item
    Human Activity Modeling on Shape Manifold
    (The Eurographics Association, 2011) Yi, Sheng; Krim, Hamid; Norris, L. K.; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    In this paper we propose a stochastic modeling of human activity on shape manifold. From a video sequence, human activity are extracted as a sequence of shape. Such sequence is considered as one realization of a random process on shape manifold. Then Different activity is modeled by manifold valued random process with different distribution. To solve the stochastic modeling on manifold, we first map the process on the shape manifold to a Euclidean process. Then the process is modeled by linear models such as stationary incremental process and a piecewise stationary incremental process. The mapping from manifold valued process to Euclidean process is known as stochastic development. The idea is to parallelly transport the tangent of curve on manifold to a single tangent space. The advantage of such technique is the one to one correspondence between the process in flat space and the one on manifold. The proposed algorithm is tested on two activity data base [RS01] [BGSB05]. The result demonstrate the high accuracy of our modeling in characterizing different activities.
  • Thumbnail Image
    Item
    Local Shape Descriptors, a Survey and Evaluation
    (The Eurographics Association, 2011) Heider, Paul; Pierre-Pierre, Alain; Li, Ruosi; Grimm, Cindy; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    Local shape descriptors can be used for a variety of tasks, from registration to comparison to shape analysis and retrieval. There have been a variety of local shape descriptors developed for these tasks, which have been evaluated in isolation or in pairs, but not against each other. We provide a survey of existing descriptors and a framework for comparing them. We perform a detailed evaluation of the descriptors using real data sets from a variety of sources. We first evaluate how stable these metrics are under changes in mesh resolution, noise, and smoothing. We then analyze the discriminatory ability of the descriptors for the task of shape matching. Our conclusion is that sampling the normal distribution and the mean curvature, using 25 samples, and reducing this data to 5-10 samples via Principal Components Analysis provides robustness to noise and the best shape discrimination results.
  • Thumbnail Image
    Item
    Partial Match of 3D Faces using Facial Curves between SIFT Keypoints
    (The Eurographics Association, 2011) Berretti, Stefano; Bimbo, Alberto Del; Pala, Pietro; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    In this work, we propose and experiment an original solution to 3D face recognition which supports partial matching of facial scans as occurs in the case of missing parts and occlusions. In the proposed approach, distinguishing traits of the face are captured by first extracting SIFT keypoints on the face scan and then measuring how the face changes along facial curves defined between pairs of keypoints. Facial curves are also associated with a measure of salience so as to distinguish curves that model characterizing traits of some subjects from curves that are frequently observed in the face of many different subjects. The recognition accuracy of the approach has been experimented on the Face Recognition Grand Challenge dataset.
  • Thumbnail Image
    Item
    Real-Time 3D Face Recognition using Line Projection and Mesh Sampling
    (The Eurographics Association, 2011) Rodrigues, Marcos A.; Robinson, A.; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    The main contribution of this paper is to present a novel method for automatic 3D face recognition based on sampling a 3D mesh structure in the presence of noise. A structured light method using line projection is employed where a 3D face is reconstructed from a single 2D shot. The process from image acquisition to recognition is described with focus on its real-time operation. Recognition results are presented and it is demonstrated that it can perform recognition in just over one second per subject in continuous operation mode and thus, suitable for real time operation.
  • Thumbnail Image
    Item
    Refining Shape Correspondence for Similar Objects Using Strain
    (The Eurographics Association, 2011) Phan, Ly; Knutsen, Andrew K.; Bayly, Philip V.; Rugonyi, Sandra; Grimm, Cindy; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    Several applications - for example, study of biological tissue movement and organ growth - require shape correspondence with a physical basis, especially for shapes or regions lacking distinctive features. For this purpose, we propose the adaptation of mechanical strain, a well-established physical measure for deformation, to the problem of constructing shape correspondence and measuring similarity between non-rigid shapes. In this paper, we demonstrate how to calculate strain for a 2D surface embedded in 3D. We then adjust the correspondence between two surfaces so that the strain varies smoothly across the deformed surface (by minimizing the change in strain). The final strain on the deformed surface can be used as a measure of shape similarity.
  • Thumbnail Image
    Item
    Selecting 3D Curves on the Nasal Surface using AdaBoost for Person Authentication
    (The Eurographics Association, 2011) Ballihi, Lahoucine; Amor, B. Ben; Daoudi, M.; Srivastava, A.; Aboutajdine, D.; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    The main contribution of this paper is the use of an AdaBoost-based learning algorithm which builds a strong classifier from a set of weak classifiers associated with level curves in the nasal region of 3D faces. Its main application is person authentication. The basic idea is to represent nasal surfaces using indexed collections of level curves, and to compare shapes of noses by comparing the shape of their corresponding curves. AdaBoost considers each curve as a weak classifier and iteratively selects relevant curves to increase the authentication accuracy. We demonstrate these ideas on a subset taken from FRGC v2 (Face Recognition Grand Challenge) database. The proposed approach increases authentication performances relative to a simple fusion of scores from all curves.
  • Thumbnail Image
    Item
    SHREC '11 Track: 3D Face Models Retrieval
    (The Eurographics Association, 2011) Veltkamp, R. C.; Jole, S. van; Drira, H.; Amor, B. Ben; Daoudi, M.; Li, H.; Chen, L.; Claes, P.; Smeets, D.; Hermans, J.; Vandermeulen, D.; Suetens, P.; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    In this paper we present the results of the 3D Shape Retrieval Contest 2011 (SHREC'11) track on face model retrieval. The aim of this track is to evaluate the performance of 3D shape retrieval algorithms that can operate on 3D face models. The benchmark dataset consists of 780 3D face scans of 130 individuals. Four groups have participated in the track with 14 method variations in total.
  • Thumbnail Image
    Item
    SHREC '11 Track: Generic Shape Retrieval
    (The Eurographics Association, 2011) Dutagaci, H.; Godil, A.; Daras, P.; Axenopoulos, A.; Litos, G.; Manolopoulou, S.; Goto, K.; Yanagimachi, T.; Kurita, Y.; Kawamura, S.; Furuya, T.; Ohbuchi, R.; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    In this paper we present the results of the 3D Shape Retrieval Contest 2011 (SHREC'11) track on generic shape retrieval. The aim of this track is to evaluate the performance of 3D shape retrieval algorithms that can operate on arbitrary 3D models. The benchmark dataset consists of 1000 3D objects classified in 50 categories. The 3D models are mainly classified based on visual shape similarity and each class has equal number of models to reduce the possible bias in evaluation results. Two groups have participated in the track with six methods in total.
  • Thumbnail Image
    Item
    SHREC '11 Track: Shape Retrieval on Non-rigid 3D Watertight Meshes
    (The Eurographics Association, 2011) Lian, Z.; Godil, A.; Bustos, B.; Daoudi, M.; Hermans, J.; Kawamura, S.; Kurita, Y.; Lavoué, G.; Nguyen, H. V.; Ohbuchi, R.; Ohkita, Y.; Ohishi, Y.; Porikli, F.; Reuter, M.; Sipiran, I.; Smeets, D.; Suetens, P.; Tabia, H.; Vandermeulen, D.; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    Non-rigid 3D shape retrieval has become an important research topic in content-based 3D object retrieval. The aim of this track is to measure and compare the performance of non-rigid 3D shape retrieval methods implemented by different participants around the world. The track is based on a new non-rigid 3D shape benchmark, which contains 600 watertight triangle meshes that are equally classified into 30 categories. In this track, 25 runs have been submitted by 9 groups and their retrieval accuracies were evaluated using 6 commonly-utilized measures.
  • Thumbnail Image
    Item
    SHREC '11: Robust Feature Detection and Description Benchmark
    (The Eurographics Association, 2011) Boyer, E.; Bronstein, A. M.; Bronstein, M. M.; Bustos, B.; Darom, T.; Horaud, R.; Hotz, I.; Keller, Y.; Keustermans, J.; Kovnatsky, A.; Litmany, R.; Reininghaus, J.; Sipiran, I.; Smeets, D.; Suetens, P.; Vandermeulen, D.; Zaharescu, A.; Zobel, V.; H. Laga and T. Schreck and A. Ferreira and A. Godil and I. Pratikakis and R. Veltkamp
    Feature-based approaches have recently become very popular in computer vision and image analysis applications, and are becoming a promising direction in shape retrieval. SHREC'11 robust feature detection and description benchmark simulates the feature detection and description stages of feature-based shape retrieval algorithms. The benchmark tests the performance of shape feature detectors and descriptors under a wide variety of transformations. The benchmark allows evaluating how algorithms cope with certain classes of transformations and strength of the transformations that can be dealt with. The present paper is a report of the SHREC'11 robust feature detection and description benchmark results