Pose Orientation Research Articles

Multi‐subregion‐based probabilistic approach to pose‐invariant face recognition

AbstractCurrent automatic facial recognition systems are not robust to changes in illumination, pose, facial expression, and occlusion. In this paper, in order to address the problem of pose change, we propose an algorithm based on a probabilistic approach to face recognition that takes into account the pose difference between probe and gallery images. By using a large facial image database, the CMU PIE database, which contains images of the same set of people taken from many different angles, we have developed a probabilistic model of how facial features change as the pose changes. This model enables us to make our face recognition system more robust to changes of pose in the probe image. The experimental results show that this approach achieves a better recognition rate than conventional face recognition methods over a much larger range of poses. For example, when the gallery contains only images of a frontal face and the probe image varies its pose orientation, the recognition rate shows less than a 10% difference until the probe pose begins to differ by more than 45°, whereas the recognition rate of a PCA‐based method begins to drop at differences as small as 10°, and that of a representative commercial system begins to drop at 30°. © 2006 Wiley Periodicals, Inc. Syst Comp Jpn, 37(8): 68–76, 2006; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.20437

Line-of-sight task-space sensing methodology for the localization of robotic end-effectors

In this paper, the implementation of a line-of-sight (LOS) task-space sensing methodology is presented for guidance-based microlocalization of robotic end-effectors. The novelty of the overall system is its applicability to cases that do not allow for the direct proximity measurement of the end-effector's pose (position and orientation). The mobility of the localization application dictates the minimum number and the type (planar or spatial) of the LOS that would be necessary to use and, consequently, the exact configuration of the sensing system. Although the main focus of the paper is the presentation of the proposed LOS sensing system, a brief discussion of a robot-guidance method, which relies on the use of this sensing system, is also included. Extensive experiments conducted for a high-precision 3-DOF (degrees of freedom) planar robotic platform utilizing the overall guidance system validated our research.

Towards a semi-automatic method for the statistically rigorous ageing of the human face

Forensic age progression for the purpose of ageing a missing child is a discipline currently dominated by artistic methodologies. In order to improve on these techniques, a statistically rigorous approach to the ageing of the human face is presented. The technique is based upon a principal component analysis and involves the definition of an ageing direction through the model space, using an age-weighted combination of model parameters. Pose and expression compensation methods are also incorporated, allowing faces at a wide variety of pose orientations and expressions to be aged accurately. Near photo-quality images are obtained quickly and the resultant ageing effects are realistic and plausible. As a quantitative check of the results, the root mean square error is calculated between the shape vector of the aged face and that of the target face, as well as between the aged face and faces of different identity at the target age. In general, this error is found to be smaller between the aged face and the target face, indicating that the face successfully retains its identity as it is aged. As a further test of the basic plausibility of our results, a regression analysis is performed between the shape model parameters and the age of each subject, assuming a linear relationship. The coefficient of determination is calculated to be r2=0.68 and the relationship between the variables is found to be significant at a level >0.99 upon performance of a standard F-test.

Multi-view face identification and pose estimation using B-spline interpolation

The available face views in the training set are mostly limited. In this paper, we present a view interpolation method using nonlinear B-spline on face manifolds. Two models, the inner–outer ellipse model and the moment of inertia model, are developed to estimate the pose orientation. We use the limited view-pose face images to form the pose eigen space. Then, based on these nonlinear manifolds we form a B-spline for each individual. Identification is to compute the shortest Euclidean distance from a given test view to the nearest point within one of these B-splines. Once the test view is classified as a familiar individual in the training set, not only can the individual be identified, but also the pose angle can be estimated. Experimental results show that B-spline interpolation can achieve a recognition rate of 95%.

Forward Problem Singularities of Manipulators Which Become PS-2RS or 2PS-RS Structures When the Actuators are Locked

The instantaneous forward problem (IFP) singularities of a parallel manipulator (PM) must be determined during the manipulator design and avoided during the manipulator operation, because they are configurations where the end-effector pose (position and orientation) cannot be controlled by acting on the actuators any longer, and the internal loads of some links become infinite. When the actuators are locked, PMs become structures consisting of one rigid body (platform) connected to another rigid body (base) by means of a number of kinematic chains (limbs). The geometries (singular geometries) of these structures where the platform can perform infinitesimal motion correspond to the IFP singularities of the PMs the structures derive from. This paper studies the singular geometries both of the PS-2RS structure and of the 2PS-RS structure. In particular, the singularity conditions of the two structures will be written in a new form. This new form will directly bring to a new geometric interpretation of their singularity conditions and to the exhaustive enumeration of the geometric conditions identifying their singular geometries. Such an enumeration will reveal that some geometric conditions were not identified in the literature. Finally, the use of the obtained results in the design of parallel manipulators which become either PS-2RS or 2PS-RS structures, when the actuators are locked, will be illustrated.

An entropic criterion for minimum uncertainty sensing in recognition and localization. II. A case study on directional distance measurements

For pt. 1 see ibid. An entropic criterion for minimum uncertainty sensing, introduced in the companion paper, is applied to a case study related to the localization and recognition of a polygonal object by means of an orientable range finder. The observed object is characterized by two different uncertain parameters: the pose (position and orientation) of the object and its identity. A priori, only partial information is available both on the object identity and on its pose. Additional information about the observed object is acquired by the orientable range finder activated according to the above criterion.

A φ-model solution for the inverse position problem of calibrated robots using virtual elementary motions

It is central to the control of manipulators to calculate the set/sets of joint-displacements which correspond to a given spatial pose (position and orientation) of the end-effector. This problem, which is referred to as the inverse position problem, represents one of the most difficult mathematical challenges in the field of robotics, particularly when performed for calibrated robots (or robots with general structures). In such cases, closed form solutions are too impractical to implement and iterative solutions suffer from numerical singularities. In the present work a procedure is introduced to obtain multiple inverse position solutions for serial robotic structures. For calibrated robots, the procedure involves a simple iterative technique designed to ensure fast convergence and eliminate the occurrence of singularity. However, inverse position solutions for spherical-wrist manipulators will be obtained in a straight-forward non-iterative fashion. A published kinematic notation, referred to as the ϕ-model, was used to develop the system equations.

Non-Contact Measurement of 6-DOF End-Effector Pose1

This paper presents a system for nan-contact measurement of 6-DOF pose (position and orientation) of a robot. The system consists of a novel sensor, three laser scanners and a computer. The sensor is rigidly attached to a robot end-effector as a short-range device (SRD). The laser scanners, fram pre-calibrated positions, generate light planes that intersect the sensor in sequential turns. These light planes are measurement references with linearly independent norms. The sensor measures light plane equations with respect to the tool coordinate frame (TCF); and the system derives 6-DOF pose of the robot from the plane equations.

Iterative multistep explicit camera calibration

Perspective camera calibration has been a research subject for a large group of researchers. However, only a small number of those methods base their approaches on the use of monoplane calibration points. We developed one of those methodologies using monoplane calibration points to perform an explicit three-dimensional (3-D) camera calibration. This methodology is based on an iterative approach. To avoid the singularities obtained with the calibration equations when monoplane calibration points are used, this method computes the calibration parameters in a multistep procedure and requires a first-guess solution for the intrinsic parameters. These intrinsic parameters are updated and their accuracy increased during the iterative procedure. All computations required are linear and in addition to the extrinsic parameters the proposed method also computes the first coefficient of the radial distortion (k/sub 1/) and the skew angle. A first-guess value for the focal length of the lens is required but its value is iteratively updated using the Gauss lens model. This methodology also includes the uncertainty horizontal image scale factor (S/sub x/) on the set of calibration parameters to be computed, which makes this approach independent of the accuracy of the horizontal scale factor. The proposed methodology has the advantage that it can be used with monoplane calibration data with no restrictions for the pose geometry of the camera. In this approach the pose estimation problem is treated separately, computing the pose orientation of the camera in a first step and using this information to compute the pose location. A model for the expected stability of the camera look angles from noisy image data and its stability analysis as a function of the pose geometry of the camera is presented. Several experimental and simulated analyzes were performed and are presented.

Pose estimation of a scanning laser Doppler vibrometer with applications to the automotive industry

A scanning laser Doppler vibrometer (SLDV) is an optical in- strument that can scan an area to measure the velocity field of struc- tures. It has been widely used in the automotive industry. In some appli- cations, the pose (position and orientation) of the SLVD are required. First, this paper describes two such applications. Then it presents a non- linear regressive model for the pose determination of the SLDV with respect to a structural coordinate system. The model is derived from coordinate transformation and the scanner model. In it, the pose is ex- pressed by six independent parameters: three for the translation vector and three for the rotation matrix. The six parameters are estimated by using the least-squares technique. Statistical inferences about this model are obtained by linear approximation. To determine the pose, four or more registration (reference) points are required. For each registration point, the structural coordinates and its corresponding input voltages need to be known. The model has been implemented in a laser-based data acquisition system, which has been used for modal analysis, struc- tural dynamic modifications, and noise control. © 1998 Society of Photo- Optical Instrumentation Engineers. (S0091-3286(98)00705-3)

Geometrical method for the determination of the position and orientation of a scanning laser doppler vibrometer

In this paper, a geometrical method is presented to determine the pose (position and orientation) of a Scanning Laser Doppler Vibrometer (SLDV) with respect to a structural coordinate system. Multiple registration points are used simultaneously in a least-squares sense. The structural coordinates and the corresponding scanning angles are known for each of the registration points. In the geometrical method three steps are involved in the determination of the SLDV pose. Its implementation has been tested by simulated data and experimental data. The results have shown that this method and its implementation are correct and effective.

An intelligent assembly robotic system based on Relative Pose measurements

One of the major problems in realizing intelligent assembly robotic systems has been the difficulty of compensating for uncertain and erroneous situations. In particular, pose (position and orientation) uncertainties lead to a challenging problem. In this paper, we outline a Relative Pose-based Strategy for effective dynamic end-pont control of manipulators in uncertain environments. The introduced strategy imposes new requirements on the higher levels of the control. Besides appropriate planning, control and sensing strategies, an advanced supervision system architecture needs to be provided. In this paper, the fundamental issues relating to intelligent assembly robotic system design are briefly discussed and an overall control architecture is introduced. The lower portions of the proposed architecture, viz. the Execution and the Supervisory Levels are described in detail.