3D Orientation Estimation Research Articles

Hierarchical Grid-based Multi-People Tracking-by-Detection With Global Optimization.

We present a hierarchical grid-based, globally optimal tracking-by-detection approach to track an unknown number of targets in complex and dense scenarios, particularly addressing the challenges of complex interaction and mutual occlusion. Frame-by-frame detection is performed by hierarchical likelihood grids, matching shape templates through a fast oriented distance transform. To allow recovery from misdetections, common heuristics such as nonmaxima suppression within observations is eschewed. Within a discretized state-space, the data association problem is formulated as a grid-based network flow model, resulting in a convex problem casted into an integer linear programming form, giving a global optimal solution. In addition, we show how a behavior cue (body orientation) can be integrated into our association affinity model, providing valuable hints for resolving ambiguities between crossing trajectories. Unlike traditional motion-based approaches, we estimate body orientation by a hybrid methodology, which combines the merits of motion-based and 3D appearance-based orientation estimation, thus being capable of dealing also with still-standing or slowly moving targets. The performance of our method is demonstrated through experiments on a large variety of benchmark video sequences, including both indoor and outdoor scenarios.

Accurate orientation estimation using AHRS under conditions of magnetic distortion.

Low cost, compact attitude heading reference systems (AHRS) are now being used to track human body movements in indoor environments by estimation of the 3D orientation of body segments. In many of these systems, heading estimation is achieved by monitoring the strength of the Earth's magnetic field. However, the Earth's magnetic field can be locally distorted due to the proximity of ferrous and/or magnetic objects. Herein, we propose a novel method for accurate 3D orientation estimation using an AHRS, comprised of an accelerometer, gyroscope and magnetometer, under conditions of magnetic field distortion. The system performs online detection and compensation for magnetic disturbances, due to, for example, the presence of ferrous objects. The magnetic distortions are detected by exploiting variations in magnetic dip angle, relative to the gravity vector, and in magnetic strength. We investigate and show the advantages of using both magnetic strength and magnetic dip angle for detecting the presence of magnetic distortions. The correction method is based on a particle filter, which performs the correction using an adaptive cost function and by adapting the variance during particle resampling, so as to place more emphasis on the results of dead reckoning of the gyroscope measurements and less on the magnetometer readings. The proposed method was tested in an indoor environment in the presence of various magnetic distortions and under various accelerations (up to 3 g). In the experiments, the proposed algorithm achieves <2° static peak-to-peak error and <5° dynamic peak-to-peak error, significantly outperforming previous methods.

Limits of 3D dipole localization and orientation estimation for single-molecule imaging: towards Green’s tensor engineering

The 3D orientation and location of individual molecules is an important marker for the local environment and the state of a molecule. Therefore dipole localization and orientation estimation is important for biological sensing and imaging. Precise dipole localization is also critical for superresolution imaging. We propose and analyze wide field microscope configurations to simultaneously measure these parameters for multiple fixed dipole emitters. Examination of the images of radiating dipoles reveals how information transfer and precise detection can be improved. We use an information theoretic analysis to quantify the performance limits of position and orientation estimation through comparison of the Cramer-Rao lower bounds in a photon limited environment. We show that bi-focal and double-helix polarization-sensitive systems are attractive candidates for simultaneously estimating the 3D dipole location and orientation.

Orientation Disparity: A Cue for 3D Orientation?

Orientation disparity, the difference in orientation that results when a texture element on a slanted surface is projected to the two eyes, has been proposed as a binocular cue for 3D orientation. Since orientation disparity is confounded with position disparity, neither behavioral nor neurophysiological experiments have successfully isolated its contribution to slant estimates or established whether the visual system uses it. Using a modified disparity energy model, we simulated a population of binocular visual cortical neurons tuned to orientation disparity and measured the amount of Fisher information contained in the activity patterns. We evaluated the potential contribution of orientation disparity to 3D orientation estimation and delimited the stimulus conditions under which it is a reliable cue. Our results suggest that orientation disparity is an efficient source of information about 3D orientation and that it is plausible that the visual system could have mechanisms that are sensitive to it. Although orientation disparity is neither necessary nor sufficient for estimating slant, it appears that it could be useful when combined with estimates from position disparity gradients and monocular perspective cues.

2D texture based classification, segmentation and 3D orientation estimation of tissues using DT-CWT feature extraction methods

In this study, four different 2D dual-tree complex wavelet (DT-CWT) based texture feature extraction methods are developed and their applications are demonstrated in segmenting and classifying tissues. Two of the methods use rotation variant texture features and the other two use rotation invariant features. This paper also proposes a novel approach to estimate 3D orientations of tissues based on rotation variant DT-CWT features. The method updates the strongest structural anisotropy direction with an iterative approach and converges to a volume orientation in few steps. Although classification and segmentation results show that there is no significant difference in the performance between rotation variant and invariant features; the latter are more robust to changes in texture rotation, which is essential for classification and segmentation of objects from 3D datasets such as medical tomography images.

Magnetic distortion in motion labs, implications for validating inertial magnetic sensors

Background Ambulatory 3D orientation estimation with Inertial Magnetic Sensor Units (IMU's) use the earth magnetic field. The magnitude of distortion in orientation in a standard equipped motion lab and its effect on the accuracy of the orientation estimation with IMU's is addressed. Methods Orientations of the earth magnetic field vectors were expressed in the laboratory's reference frame. The effect of a distorted earth magnetic field on orientation estimation with IMU's (using both a quaternion and a Kalman fusing algorithm) was compared to orientations derived from an optical system. Findings The magnetic field varied considerably, with the strongest effects at 5 cm above floor level with a standard deviation in heading of 29°, decreasing to 3° at levels higher than 100 cm. Orientation estimation was poor with the quaternion filter, for the Kalman filter results were acceptable, despite a systematic deterioration over time (after 20–30 s). Interpretation Distortion of the earth magnetic field is depending on construction materials used in the building, and should be taken into account for calibration, alignment to a reference system, and further measurements. Mapping the measurement volume to determine its ferromagnetic characteristics in advance of planned experiments can be the rescue of the data set. Conclusions To obtain valid data, “mapping” of the laboratory is essential, although less critical with the Kalman filter and at larger distances (>100 cm) from suspect materials. Measurements should start in a “safe” area and continue no longer than 20–30 s in a heavily distorted earth magnetic field.

Tracking a detected face with dynamic programming

In this paper, we consider the problem of tracking a moving human face in front of a video camera in real-time for a model-based coding application. The 3D head tracking in a MBC system could be implemented sequentially as 2D location tracking, coarse 3D orientation estimation and accurate 3D motion estimation. This work focuses on the 2D location tracking of one face object through continuously using a face detector. The face detection scheme is based on a boosted cascade of simple Haar-like feature classifiers. Although such a detector demonstrated rapid processing speed, high detection rate can only be achieved for rather strictly near front faces. This introduces the ‘loss of tracking’ problem when used in 2D tracking. This paper suggests an easy method of solving the pose problem by using the technique of Dynamic Programming. The Haar-like facial features used in the 2D face detector are spatially arranged into a 1D deformable face graph and the Dynamic Programming matching is used to handle the ‘loss of track’ problem. Dynamic Programming matches the deformed version of the face graph extracted from a rotated face with the template taken online before ‘loss of tracking’ happens. Since the deformable face graph covers a big pose variation, the developed technique is robust in tracking rotated faces. Embedding Haar-like facial features into a deformable face graph is the key feature of our tracking scheme. A real time tracking system based on this technique has been set up and tested. Encouraging results have been got and are reported.

Orientation in Manhattan: equiprojective classes and sequential estimation

The problem of inferring 3D orientation of a camera from video sequences has been mostly addressed by first computing correspondences of image features. This intermediate step is now seen as the main bottleneck of those approaches. In this paper, we propose a new 3D orientation estimation method for urban (indoor and outdoor) environments, which avoids correspondences between frames. The scene property exploited by our method is that many edges are oriented along three orthogonal directions; this is the recently introduced Manhattan world (MW) assumption. The main contributions of this paper are: the definition of equivalence classes of equiprojective orientations, the introduction of a new small rotation model, formalizing the fact that the camera moves smoothly, and the decoupling of elevation and twist angle estimation from that of the compass angle. We build a probabilistic sequential orientation estimation method, based on an MW likelihood model, with the above-listed contributions allowing a drastic reduction of the search space for each orientation estimate. We demonstrate the performance of our method using real video sequences.

Three-dimensional location estimation of circular features for machine vision

A closed-form analytical solution to the problem of 3D estimation of circular-feature location is presented. Two different cases are considered: 3D orientation and position estimation when the radius is known and when it is not known. Extension of the method to general 3D quadratic features is also addressed. Simulated experimental results obtained for all three cases verified the analytical method. In the case of real experiments, a set of circles located on a calibration plate, whose locations were known with respect to a reference frame, were used for camera calibration as well as for the application of the method. A sequential compensation procedure was applied to the input gray-level image to compensate for distortion. These results also showed the validity of the process and the applicability of the analytical method.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>