Showing papers on "Pose published in 2003"

Fast pose estimation with parameter-sensitive hashing

Abstract: Example-based methods are effective for parameter estimation problems when the underlying system is simple or the dimensionality of the input is low. For complex and high-dimensional problems such as pose estimation, the number of required examples and the computational complexity rapidly become prohibitively high. We introduce a new algorithm that learns a set of hashing functions that efficiently index examples relevant to a particular estimation task. Our algorithm extends locality-sensitive hashing, a recently developed method to find approximate neighbors in time sublinear in the number of examples. This method depends critically on the choice of hash functions that are optimally relevant to a particular estimation problem. Experiments demonstrate that the resulting algorithm, which we call parameter-sensitive hashing, can rapidly and accurately estimate the articulated pose of human figures from a large database of example images.

Linear pose estimation from points or lines

Abstract: Estimation of camera pose from an image of n points or lines with known correspondence is a thoroughly studied problem in computer vision. Most solutions are iterative and depend on nonlinear optimization of some geometric constraint, either on the world coordinates or on the projections to the image plane. For real-time applications, we are interested in linear or closed-form solutions free of initialization. We present a general framework which allows for a novel set of linear solutions to the pose estimation problem for both n points and n lines. We then analyze the sensitivity of our solutions to image noise and show that the sensitivity analysis can be used as a conservative predictor of error for our algorithms. We present a number of simulations which compare our results to two other recent linear algorithms, as well as to iterative approaches. We conclude with tests on real imagery in an augmented reality setup.

Estimating 3D hand pose from a cluttered image

Abstract: A method is proposed that can generate a ranked list of plausible three-dimensional hand configurations that best match an input image. Hand pose estimation is formulated as an image database indexing problem, where the closest matches for an input hand image are retrieved from a large database of synthetic hand images. In contrast to previous approaches, the system can function in the presence of clutter, thanks to two novel clutter-tolerant indexing methods. First, a computationally efficient approximation of the image-to-model chamfer distance is obtained by embedding binary edge images into a high-dimensional Euclidean space. Second, a general-purpose, probabilistic line matching method identifies those line segment correspondences between model and input images that are the least likely to have occurred by chance. The performance of this clutter tolerant approach is demonstrated in quantitative experiments with hundreds of real hand images.

A real-time tracker for markerless augmented reality

Abstract: Augmented reality has now progressed to the point where real-time applications are required and being considered. At the same time it is important that synthetic elements are rendered and aligned in the scene in an accurate and visually acceptable way. In order to address these issues a real-time, robust and efficient 3D model-based tracking algorithm is proposed for a 'video see through' monocular vision system. The tracking of objects in the scene amounts to calculating the pose between the camera and the objects. Virtual objects can then be projected into the scene using the pose. Here, non-linear pose computation is formulated by means of a virtual visual servoing approach. In this context, the derivation of point-to-curve interaction matrices is given for different features including lines, circles, cylinders and spheres. A local moving edge tracker is used in order to provide real-time tracking of points normal to the object contours. A method is proposed for combining local position uncertainty and global pose uncertainty in an efficient and accurate way by propagating uncertainty. Robustness is obtained by integrating an M-estimator into the visual control law via an iteratively re-weighted least squares implementation. The method presented in this paper has been validated on several complex image sequences including outdoor environments. Results show the method to be robust to occlusion, changes in illumination and mistracking.

Single-camera tracking of an object

Abstract: A method and system for determining the position and orientation of an object (150) is disclosed. A set of markers (104) attached or associated with the object (150) is optically tracked and geometric translation is performed (110) to use the coordinates of the set of markers (104) to determine the location and orientation of their associated object (150).

Pose estimation using line-based dynamic vision and inertial sensors

Abstract: An observer problem from a computer vision application is studied. Rigid body pose estimation using inertial sensors and a monocular camera is considered and it is shown how rotation estimation can be decoupled from position estimation. Orientation estimation is formulated as an observer problem with implicit output where the states evolve on SO(3). A careful observability study reveals interesting group theoretic structures tied to the underlying system structure. A locally convergent observer where the states evolve on SO (3) is proposed and numerical estimates of the domain of attraction is given. Further, it is shown that, given convergent orientation estimates, position estimation can be formulated as a linear implicit output problem. From an applications perspective, it is outlined how delayed low bandwidth visual observations and high bandwidth rate gyro measurements can provide high bandwidth estimates. This is consistent with real-time constraints due to the complementary characteristics of the sensors which are fused in a multirate way.

LADAR-based detection and tracking of moving objects from a ground vehicle at high speeds

Abstract: Detection and tracking of moving objects (DATMO) in crowded urban areas from a ground vehicle at high speeds is difficult because of a wide variety of targets and uncertain pose estimation from odometry and GPS/DGPS. In this paper we present a solution of the simultaneous localization and mapping (SLAM) with DATMO problem to accomplish this task using ladar sensors and odometry. With a precise pose estimate and a surrounding map from SLAM, moving objects are detected without a priori knowledge of the targets. The interacting multiple model (IMM) estimation algorithm is used for modeling the motion of a moving object and to predict its future location. The multiple hypothesis tracking (MHT) method is applied to refine detection and data association. Experimental results demonstrate that our algorithm is reliable and robust to detect and track pedestrians and different types of moving vehicles in urban areas.

Adaptive view-based appearance models

Abstract: We present a method for online rigid object tracking using an adaptive view-based appearance model. When the object's pose trajectory crosses itself, our tracker has bounded drift and can track objects undergoing large motion for long periods of time. Our tracker registers each incoming frame against the views of the appearance model using a two-frame registration algorithm. Using a linear Gaussian filter, we simultaneously estimate the pose of the object and adjust the view-based model as pose-changes are recovered from the registration algorithm. The adaptive view-based model is populated online with views of the object as it undergoes different orientations in pose space, allowing us to capture non-Lambertian effects. We tested our approach on a real-time rigid object tracking task using stereo cameras and observed an RMS error within the accuracy limit of an attached inertial sensor.

A hand-pose estimation for vision-based human interfaces

Abstract: This paper proposes a novel method for a hand-pose estimation that can be used for vision-based human interfaces. The aim of this method is to estimate all joint angles. In this method, the hand regions are extracted from multiple images obtained by a multiviewpoint camera system. By integrating these multiviewpoint silhouette images, a hand pose is reconstructed as a "voxel model." Then, all joint angles are estimated using a three-dimensional model fitting between the hand model and the voxel model. The following two experiments were performed: (1) an estimation of joint angles by the silhouette images from the hand-pose simulator and (2) hand-pose estimation using real hand images. The experimental results indicate the feasibility of the proposed algorithm for vision-based interfaces, although the algorithm requires faster implementation for real-time processing.

Attractive People: Assembling Loose-Limbed Models using Non-parametric Belief Propagation

Abstract: The detection and pose estimation of people in images and video is made challenging by the variability of human appearance, the complexity of natural scenes, and the high dimensionality of articulated body models. To cope with these problems we represent the 3D human body as a graphical model in which the relationships between the body parts are represented by conditional probability distributions. We formulate the pose estimation problem as one of probabilistic inference over a graphical model where the random variables correspond to the individual limb parameters (position and orientation). Because the limbs are described by 6-dimensional vectors encoding pose in 3-space, discretization is impractical and the random variables in our model must be continuous-valued. To approximate belief propagation in such a graph we exploit a recently introduced generalization of the particle filter. This framework facilitates the automatic initialization of the body-model from low level cues and is robust to occlusion of body parts and scene clutter.

Method and apparatus for single image 3D vision guided robotics

Abstract: A method of three-dimensional object location and guidance to allow robotic manipulation of an object with variable position and orientation using a sensor array which is a collection of one or more sensors capable of forming a single image.

Driver's view and vehicle surround estimation using omnidirectional video stream

Abstract: Our research is focused on the development of novel machine vision based telematic systems, which provide non-intrusive probing of the state of the driver and driving conditions. In this paper we present a system which allows simultaneous capture of the driver's head pose, driving view, and surroundings of the vehicle. The integrated machine vision system utilizes a video stream of full 360 degree panoramic field of view. The processing modules include perspective transformation, feature extraction, head detection, head pose estimation, driving view synthesis, and motion segmentation. The paper presents a multi-state statistical decision models with Kalman filtering based tracking for head pose detection and face orientation estimation. The basic feasibility and robustness of the approach is demonstrated with a series of systematic experimental studies.

Moving object detection device, moving object detection method, and moving object detection program

Abstract: A moving object detection device comprising: an object distance setting part, determining the distance to a moving object that moves the most based on a distance image, in which information on distances to image-taken objects are embedded, and a difference image, in which the movements of moving objects are embedded as movement information; an object distance image generating part, generating an object distance image corresponding to the abovementioned distance; and a contour extraction part, extracting a contour inside the object distance image to detect a moving object.

Avoiding replay-attacks in a face recognition system using head-pose estimation

Abstract: The subject is a challenge response mechanism used as an optional add-on to the face recognition part of the multi modal biometric authentication system "BioID" This mechanism greatly enhances security in regard to replay attacks The user is required to look into a certain direction, which is randomly chosen by the system By estimating the head pose, the system verifies the user's response to the direction challenge The pose estimation is based on detection and subsequent tracking of suitable facial features Experimental evaluations have shown the effectiveness of the approach against replay attacks

Point fingerprint: a new 3-D object representation scheme

Abstract: This paper proposes a new, efficient surface representation method for surface matching. A feature carrier for a surface point, which is a set of two-dimensional (2-D) contours that are the projections of geodesic circles on the tangent plane, is generated. The carrier is named point fingerprint because its pattern is similar to human fingerprints and plays a role in discriminating surface points. Corresponding points on surfaces from different views are found by comparing their fingerprints. The point fingerprint is able to carry curvature, color, and other information which can improve matching accuracy, and the matching process is faster than 2-D image comparison. A novel candidate point selection method based on the fingerprint irregularity is introduced. Point fingerprint is successfully applied to pose estimation of real range data.

Head pose estimation using Fisher Manifold learning

Abstract: Here, we propose a new learning strategy for head pose estimation. Our approach uses nonlinear interpolation to estimate the head pose using the learning result from face images of two head poses. Advantage of our method to regression method is that it only requires training images of two head poses and better generalization ability. It outperforms existed methods, such as regression and multiclass classification method, on both synthesis and real face images. Average head pose estimation error of yaw rotation is about 4/sup 0/, which proves that our method is effective in head pose estimation.

Relative Pose Estimation for Instrumented, Calibrated Imaging Platforms

Abstract: Recent efforts in robust estimation of the two-view relation have fo- cused on uncalibrated cameras with no prior knowledge of pose. However, in practice robotic vehicles that perform image-based navigation and mapping typi- cally do carry a calibrated camera and pose sensors; this additional knowledge is currently not being exploited. This paper presents three contributions in using vision with instrumented and cal- ibrated platforms. First, we improve the performace of the correspondence stage by using uncertain measurements from egomotion sensors to constrain possible matches. Second, we assume wide-baseline conditions and propose Zernike mo- ments to describe affine invariant features. Third, we robustly estimate the essen- tial matrix with a new 6-point algorithm. Our solution is simpler than the minimal 5-point one and, unlike the linear 6-point solution, does not fail on planar scenes. While the contributions are general, we present structure and motion results from an underwater robotic survey.

Real-time tracking and pose estimation for industrial objects using geometric features

Abstract: This paper presents a fast tracking algorithm capable of estimating the complete pose (6DOF) of an industrial object by using its circular-shape features. Since the algorithm is part of a real-time visual servoing system designed for assembly of automotive parts on-the-fly, the main constraints in the design of the algorithm were: speed and accuracy. That is: close to frame-rate performance, and error in pose estimation smaller than a few millimeters. The algorithm proposed uses only three model features, and yet it is very accurate and robust. For that reason both constraints were satisfied: the algorithm runs at 60 fps (30 fps for each stereo image) on a PIII-800 MHz computer, and the pose of the object is calculated within an uncertainty of 2.4 mm in translation and 1.5 degree in rotation.

First contact: an active vision approach to segmentation

Abstract: How a robot should grasp an object depends on its size and shape. Such parameters can be estimated visually, but this is fallible, particularly for unrecognized, unfamiliar objects. Failure will result in a clumsy grasp or glancing blow against the object. If the robot does not learn something from the encounter, then it will be apt to repeat the same mistake again and again. This paper shows how to recover information about an object's extent by poking it, either accidentally or deliberately. Poking an object makes it move, and motion is a powerful cue for visual segmentation. The periods immediately before and after the moment of impact turn out to be particularly informative, and give visual evidence for the boundary of the object that is well suited to segmentation using graph cuts. The segmentation algorithm is shown to produce results consistent enough to support autonomous collection of datasets for object recognition, which enables often-encountered objects to be segmented without the need for further poking.

Pose estimation using 3D view-based eigenspaces

Abstract: We present a method for estimating the absolute pose of a rigid object based on intensity and depth view-based eigenspaces, built across multiple views of example objects of the same class. Given an initial frame of an object with unknown pose, we reconstruct a prior model for all views represented in the eigenspaces. For each new frame, we compute the pose-changes between every view of the reconstructed prior model and the new frame. The resulting pose-changes are then combined and used in a Kalman filter update. This approach for pose estimation is user-independent and the prior model can be initialized automatically from any viewpoint of the view-based eigenspaces. To track more robustly over time, we present an extension of this pose estimation technique where we integrate our prior model approach with an adaptive differential tracker. We demonstrate the accuracy of our approach on face pose tracking using stereo cameras.

Human body pose estimation using silhouette shape analysis

Abstract: We describe a system for human body pose estimation from multiple views that is fast and completely automatic. The algorithm works in the presence of multiple people by decoupling the problems of pose estimation of different people. The pose is estimated based on a likelihood function that integrates information from multiple views and thus obtains a globally optimal solution. Other characteristics that make our method more general than previous work include: (1) no manual initialization; (2) no specification of the dimensions of the 3D structure; (3) no reliance on some learned poses or patterns of activity; (4) insensitivity to edges and clutter in the background and within the foreground. The algorithm has applications in surveillance and promising results have been obtained.

A faster method for 3D/2D medical image registration—a simulation study

Abstract: 3D/2D patient-to-computed-tomography (CT) registration is a method to determine a transformation that maps two coordinate systems by comparing a projection image rendered from CT to a real projection image. Iterative variation of the CT's position between rendering steps finally leads to exact registration. Applications include exact patient positioning in radiation therapy, calibration of surgical robots, and pose estimation in computer-aided surgery. One of the problems associated with 3D/2D registration is the fact that finding a registration includes solving a minimization problem in six degrees of freedom (dof) in motion. This results in considerable time requirements since for each iteration step at least one volume rendering has to be computed. We show that by choosing an appropriate world coordinate system and by applying a 2D/2D registration method in each iteration step, the number of iterations can be grossly reduced from n6 to n5. Here, n is the number of discrete variations around a given coordinate. Depending on the configuration of the optimization algorithm, this reduces the total number of iterations necessary to at least 1/3 of it's original value. The method was implemented and extensively tested on simulated x-ray images of a tibia, a pelvis and a skull base. When using one projective image and a discrete full parameter space search for solving the optimization problem, average accuracy was found to be 1.0 ± 0.6(°) and 4.1 ± 1.9 (mm) for a registration in six parameters, and 1.0 ± 0.7(°) and 4.2 ± 1.6 (mm) when using the 5 + 1 dof method described in this paper. Time requirements were reduced by a factor 3.1. We conclude that this hardware-independent optimization of 3D/2D registration is a step towards increasing the acceptance of this promising method for a wide number of clinical applications.

Dense shape reconstruction of a moving object under arbitrary, unknown lighting

Abstract: We present a method for shape reconstruction from several images of a moving object. The reconstruction is dense (up to image resolution). The method assumes that the motion is known, e.g., by tracking a small number of feature points on the object. The object is assumed Lambertian (completely matte), light sources should not be very close to the object but otherwise arbitrary, and no knowledge of lighting conditions is required. An object changes its appearance significantly when it changes its orientation relative to light sources, causing violation of the common brightness constancy assumption. While a lot of effort is devoted to deal with this violation, we demonstrate how to exploit it to recover 3D structure from 2D images. We propose a new correspondence measure that enables point matching across views of a moving object. The method has been tested both on computer simulated examples and on a real object.

Outlier modeling in image matching

Abstract: We address the question of how to characterize the outliers that may appear when matching two views of the same scene. The match is performed by comparing the difference of the two views at a pixel level aiming at a better registration of the images. When using digital photographs as input, we notice that an outlier is often a region that has been occluded, an object that suddenly appears in one of the images, or a region that undergoes an unexpected motion. By assuming that the error in pixel intensity generated by the outlier is similar to an error generated by comparing two random regions in the scene, we can build a model for the outliers based on the content of the two views. We illustrate our model by solving a pose estimation problem: the goal is to compute the camera motion between two views. The matching is expressed as a mixture of inliers versus outliers, and defines a function to minimize for improving the pose estimation. Our model has two benefits: First, it delivers a probability for each pixel to belong to the outliers. Second, our tests show that the method is substantially more robust than traditional robust estimators (M-estimators) used in image stitching applications, with only a slightly higher computational complexity.

Pose estimation of cylindrical fragments for semi-automatic bone fracture reduction

Abstract: We present an approach for estimating the relative transformations between fragments of a broken cylindrical structure in 3d. To solve this problem, we first measure the orientation and position of the cylinder axes for each fragment by an adapted kind of Hough Transformation. The cylinder axes are an important feature for separation of fractured areas and for calculation of an initial reposition solution (constraining 4 DOFs). After these processing steps, we compute the relative transformations between corresponding fragments by using well-known surface registration techniques, like 2d depths correlation and the ICP (Iterative Closest Point) algorithm. One goal of our project is to use the proposed method for estimation of relative transformations between fragments of fractured long bones for computer aided and semi-automatic bone alignment and fracture reduction in surgery.

A graphics hardware implementation of the generalized Hough transform for fast object recognition, scale, and 3D pose detection

Abstract: The generalized Hough transform constitutes a wellknown approach to object recognition and pose detection. To attain reliable detection results, however, a very large number of candidate object poses and scale factors need to be considered. We employ an inexpensive, consumer-market graphics-card as the "poor man's" parallel processing system. We describe the implementation of a fast and enhanced version of the generalized Hough transform on graphics hardware. Thanks to the high bandwidth of on-board texture memory, a single pose can be evaluated in less than 3 ms, independent of the number of edge pixels in the image. From known object geometry, our hardware-accelerated generalized Hough transform algorithm is capable of detecting an object's 3D pose, scale, and position in the image within less than one minute. A good pose estimation is even delivered in less than 10 seconds.

Absolute head pose estimation from overhead wide-angle cameras

Abstract: Most surveillance cameras have a wide-angle field of view and are situated unobtrusively at overhead positions. For this type of application, head pose estimation is very challenging because of the limitations of the quality and resolution of the incoming data. In addition, even though the absolute head pose is constant, the head pose in camera view changes depending upon the location of head with respect the camera. We present a solution to estimate absolute coarse head pose for wide-angle overhead cameras by integrating 3D head position and pose information. The work involves image-based learning, pose correction based on 3D position, and real-time multicamera integration of low-resolution imagery. The system can be applied to an active face catalogger to obtain the best view of the face for surveillance, to customer relationship management to record behavior in retail stores or to virtual reality as an input device.

A complete symbolic-numeric linear method for camera pose determination

Abstract: Camera pose estimation is the problem of determining the position and orientation of an internally calibrated camera from known 3D reference points and their images We briefly survey several existing methods for pose estimation, then introduce our new complete linear method, which is based on a symbolic-numeric method from the geometric (Jet) theory of partial differential equations The method is stable and robust In particular, it can deal with the points near critical configurations Numerical experiments are given to show the performance of the new method