Results 1  10
of
14
Human action recognition by representing 3d skeletons as points in a lie group. In
 CVPR,
, 2014
"... Abstract Recently introduced costeffective depth sensors coupled with the realtime skeleton estimation algorithm of Shotton et al. ..."
Abstract

Cited by 16 (0 self)
 Add to MetaCart
(Show Context)
Abstract Recently introduced costeffective depth sensors coupled with the realtime skeleton estimation algorithm of Shotton et al.
From Manifold to Manifold: GeometryAware Dimensionality Reduction for SPD Matrices
"... of any given curve under the geodesic distance δg and the Stein metric δS up to scale of 2 √ 2. The proof of this theorem follows several steps. We start with the definition of curve length and intrinsic metric. Without any assumption on differentiability, let (M, d) be a metric space. A curve in M ..."
Abstract

Cited by 8 (1 self)
 Add to MetaCart
(Show Context)
of any given curve under the geodesic distance δg and the Stein metric δS up to scale of 2 √ 2. The proof of this theorem follows several steps. We start with the definition of curve length and intrinsic metric. Without any assumption on differentiability, let (M, d) be a metric space. A curve in M is a continuous function γ: [0, 1] → M and joins the starting point γ(0) = x to the end point γ(1) = y. Definition 1. The length of a curve γ is the supremum of l(γ; {ti}) over all possible partitions {ti}, where 0 = t0 < t1 < · · · < tn−1 < tn = 1 and l(γ; {ti}) = ∑ i d (γ(ti), γ(ti−1)). Definition 2. The intrinsic metric ̂ δ(x, y) on M is defined as the infimum of the lengths of all paths from x to y. Theorem 1 ( [2]). If the intrinsic metrics induced by two metrics d1 and d2 are identical up to a scale ξ, then the length of any given curve is the same under both metrics up to ξ. Theorem 2 ( [2]). If d1(x, y) and d2(x, y) are two metrics defined on a space M such that d2(x, y) lim = 1. (1) d1(x,y)→0 d1(x, y) uniformly (with respect to x and y), then their intrinsic metrics are identical. Therefore, here, we need to study the behavior of lim δ 2 S (X,Y)→0 δ 2 g(X, Y) δ2 S
Bregman Divergences for Infinite Dimensional Covariance Matrices
"... We introduce an approach to computing and comparing Covariance Descriptors (CovDs) in infinitedimensional spaces. CovDs have become increasingly popular to address classification problems in computer vision. While CovDs offer some robustness to measurement variations, they also throw away part of t ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
(Show Context)
We introduce an approach to computing and comparing Covariance Descriptors (CovDs) in infinitedimensional spaces. CovDs have become increasingly popular to address classification problems in computer vision. While CovDs offer some robustness to measurement variations, they also throw away part of the information contained in the original data by only retaining the secondorder statistics over the measurements. Here, we propose to overcome this limitation by first mapping the original data to a highdimensional Hilbert space, and only then compute the CovDs. We show that several Bregman divergences can be computed between the resulting CovDs in Hilbert space via the use of kernels. We then exploit these divergences for classification purpose. Our experiments demonstrate the benefits of our approach on several tasks, such as material and texture recognition, person reidentification, and action recognition from motion capture data. 1.
Sparse Coding on Symmetric Positive Definite Manifolds using Bregman Divergences
"... Abstract—This paper introduces sparse coding and dictionary learning for Symmetric Positive Definite (SPD) matrices, which are often used in machine learning, computer vision and related areas. Unlike traditional sparse coding schemes that work in vector spaces, in this paper we discuss how SPD matr ..."
Abstract

Cited by 2 (2 self)
 Add to MetaCart
(Show Context)
Abstract—This paper introduces sparse coding and dictionary learning for Symmetric Positive Definite (SPD) matrices, which are often used in machine learning, computer vision and related areas. Unlike traditional sparse coding schemes that work in vector spaces, in this paper we discuss how SPD matrices can be described by sparse combination of dictionary atoms, where the atoms are also SPD matrices. We propose to seek sparse coding by embedding the space of SPD matrices into Hilbert spaces through two types of Bregman matrix divergences. This not only leads to an efficient way of performing sparse coding, but also an online and iterative scheme for dictionary learning. We apply the proposed methods to several computer vision tasks where images are represented by region covariance matrices. Our proposed algorithms outperform stateoftheart methods on a wide range of classification tasks, including face recognition, action recognition, material classification and texture categorization. Index Terms—Riemannian geometry, Bregman divergences, kernel methods, sparse coding, dictionary learning. I.
Bregman Divergences for Infinite Dimensional Covariance Matrices
"... Abstract We introduce an approach to computing and comparing Covariance Descriptors (CovDs) ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
Abstract We introduce an approach to computing and comparing Covariance Descriptors (CovDs)
From Manifold to Manifold: GeometryAware Dimensionality Reduction for SPD Matrices
"... Abstract. Representing images and videos with Symmetric Positive Definite (SPD) matrices and considering the Riemannian geometry of the resulting space has proven beneficial for many recognition tasks. Unfortunately, computation on the Riemannian manifold of SPD matrices especially of highdimensi ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
Abstract. Representing images and videos with Symmetric Positive Definite (SPD) matrices and considering the Riemannian geometry of the resulting space has proven beneficial for many recognition tasks. Unfortunately, computation on the Riemannian manifold of SPD matrices especially of highdimensional onescomes at a high cost that limits the applicability of existing techniques. In this paper we introduce an approach that lets us handle highdimensional SPD matrices by constructing a lowerdimensional, more discriminative SPD manifold. To this end, we model the mapping from the highdimensional SPD manifold to the lowdimensional one with an orthonormal projection. In particular, we search for a projection that yields a lowdimensional manifold with maximum discriminative power encoded via an affinityweighted similarity measure based on metrics on the manifold. Learning can then be expressed as an optimization problem on a Grassmann manifold. Our evaluation on several classification tasks shows that our approach leads to a significant accuracy gain over stateoftheart methods.
The value of multiple viewpoints in gesturebased user authentication
 in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern RecognitionWorkshop
, 2014
"... Although traditionally used as a gesture recognition device, the Kinect has been recently leveraged for user entry control. In this context, a user admission decision is typically based on biometrics such as face, speech, gait and gestures. Despite being a relatively new biometric, gestures have b ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
(Show Context)
Although traditionally used as a gesture recognition device, the Kinect has been recently leveraged for user entry control. In this context, a user admission decision is typically based on biometrics such as face, speech, gait and gestures. Despite being a relatively new biometric, gestures have been shown to be a promising authentication modality. These results have been achieved using a single Kinect camera. This paper aims to investigate the potential performance and robustness gains in gesturebased user authentication using multiple Kinects. We study the impact of multiple viewpoints on a dataset of 40 users that contains notable degradations from user memory and personal effects (multiple types of bags and outerwear). We found that two additional viewpoints can provide as much as 26–43 % average relative improvement in the Equal Error Rate (EER) for user authentication, and as much as 16–68 % average relative improvement in the Correct Classification Error (CCE) compared to using a single centered Kinect camera. 1.
Mining Key Skeleton Poses with Latent SVM for Action Recognition
"... Human action recognition based on 3D skeleton has become an active research field in recent years with the recently developed commodity depth sensors. Most published methods analyze an entire 3D depth data, construct midlevel part representations, or use trajectory descriptor of spatialtemporal i ..."
Abstract
 Add to MetaCart
(Show Context)
Human action recognition based on 3D skeleton has become an active research field in recent years with the recently developed commodity depth sensors. Most published methods analyze an entire 3D depth data, construct midlevel part representations, or use trajectory descriptor of spatialtemporal interest point for recognizing human activities. Unlike previous work, a novel and simple action representation is proposed in this paper which models the action as a sequence of inconsecutive and discriminative skeleton poses, named as key skeleton poses. The pairwise relative positions of skeleton joints are used as feature of the skeleton poses which are mined with the aid of the latent support vector machine (latent SVM). The advantage of our method is resisting against intraclass variation such as noise and large nonlinear temporal deformation of human action. We evaluate the proposed approach on three benchmark action datasets captured by Kinect devices: MSR Action 3D dataset, UTKinect Action dataset, and Florence 3D Action dataset. The detailed experimental results demonstrate that the proposed approach achieves superior performance to the stateoftheart skeletonbased action recognition methods.
Jointly Learning Heterogeneous Features for RGBD Activity Recognition JianFang Hu†, WeiShi Zheng‡⋆
"... In this paper, we focus on heterogeneous feature learning for RGBD activity recognition. Considering that features from different channels could share some similar hidden structures, we propose a joint learning model to simultaneously explore the shared and featurespecific components as an instan ..."
Abstract
 Add to MetaCart
(Show Context)
In this paper, we focus on heterogeneous feature learning for RGBD activity recognition. Considering that features from different channels could share some similar hidden structures, we propose a joint learning model to simultaneously explore the shared and featurespecific components as an instance of heterogenous multitask learning. The proposed model in an unified framework is capable of: 1) jointly mining a set of subspaces with the same dimensionality to enable the multitask classifier learning, and 2) meanwhile, quantifying the shared and featurespecific components of features in the subspaces. To efficiently train the joint model, a threestep iterative optimization algorithm is proposed, followed by two inference models. Extensive results on three activity datasets have demonstrated the efficacy of the proposed method. In addition, a novel RGBD activity dataset focusing on humanobject interaction is collected for evaluating the proposed method, which will be made available to the community for RGBD activity benchmarking and analysis. 1.
BioInspired Predictive Orientation Decomposition of Skeleton Trajectories for RealTime Human Activity Prediction
"... Abstract — Activity prediction is an essential task in practical humancentered robotics applications, such as security, assisted living, etc., which targets at inferring ongoing human activities based on incomplete observations. To address this challenging problem, we introduce a novel bioinspired ..."
Abstract
 Add to MetaCart
(Show Context)
Abstract — Activity prediction is an essential task in practical humancentered robotics applications, such as security, assisted living, etc., which targets at inferring ongoing human activities based on incomplete observations. To address this challenging problem, we introduce a novel bioinspired predictive orientation decomposition (BIPOD) approach to construct representations of people from 3D skeleton trajectories. Our approach is inspired by biological research in human anatomy. In order to capture spatiotemporal information of human motions, we spatially decompose 3D human skeleton trajectories and project them onto three anatomical planes (i.e., coronal, transverse and sagittal planes); then, we describe shortterm time information of joint motions and encode highorder temporal dependencies. By estimating future skeleton trajectories that are not currently observed, we endow our BIPOD representation with the critical predictive capability. Empirical studies validate that our BIPOD approach obtains promising performance, in terms of accuracy and efficiency, using a physical TurtleBot2 robotic platform to recognize ongoing human activities. Experiments on benchmark datasets further demonstrate that our new BIPOD representation significantly outperforms previous approaches for realtime activity classification and prediction from 3D human skeleton trajectories. I.