Object Articulation Based on Local 3D Motion Estimation

If 3D rigid motion is estimated with some error a distorted version of the scene structure will in turn be computed. Of computational interest are these regions in space where the distortions are such that the depths become negative, because in order to be visible the scene has to lie in front of the image. The stability analysis for the structure-from-motion problem presented in this paper investigates the optimal relationship between the errors in the estimated translational and rotational parameters of a rigid motion, that results in the estimation of a minimum number of negative depth values. The input used is the value of the flow along some direction, which is more general than optic flow or correspondence. For a planar retina it is shown that the optimal configuration is achieved when the projections of the translational and rotational errors on the image plane are perpendicular. Furthermore, the projection of the actual and the estimated translation lie on a line passing through the image center. For a spherical retina given a rotational error, the optimal translation is the correct one, while given a translational error. The optimal rotational error is normal to the translational one at an equal distance from the real and estimated translations. The proofs, besides illuminating the confounding of translation and rotation in structure from motion, have an important application to ecological optics, explaining differences of planar and spherical eye or camera designs in motion and shape estimation.

If 3D rigid motion can be correctly estimated from image sequences, the structure of the scene can be correctly derived using the equations for image formation. However, an error in the estimation of 3D motion will result in the computation of a distorted version of the scene structure. Of computational interest are these regions in space where the distortions are such that the depths become negative, because in order for the scene to be visible it has to lie in front of the image, and thus the corresponding depth estimates have to be positive. The stability analysis for the structure from motion problem presented in this paper investigates the optimal relationship between the errors in the estimated translational and rotational parameters of a rigid motion that results in the estimation of a minimum number of negative depth values. The input used is the value of the flow along some direction, which is more general than optic flow or correspondence. For a planar retina it is shown that the optimal configuration is achieved when the projections of the translational and rotational errors on the image plane are perpendicular. Furthermore, the projection of the actual and the estimated translation lie on a line through the center. For a spherical retina, given a rotational error, the optimal translation is the correct one; given a translational error, the optimal rotational negative deptherror depends both in direction and value on the actual and estimated translation as well as the scene in view. The proofs, besides illuminating the confounding of translation and rotation in structure from motion, have an important application to ecological optics. The same analysis provides a computational explanation of why it is easier to estimate self-motion in the case of a spherical retina and why shape can be estimated easily in the case of a planar retina, thus suggesting that nature‘s design of compound eyes (or panoramic vision) for flying systems and camera-type eyes for primates (and other systems that perform manipulation) is optimal.

Self-supervised rigid object 3-D motion estimation from monocular video

Proposes a robust pose estimation algorithm from 2D correspondences, which is a key issue of a 3D face modeling system from calibrated stereo sequences. The estimated rigid motion parameters are utilized to obtain the perspective projection of a generic face model, which is then matched with 2D clues extracted from the image under corresponding pose to decide the shape of a specified face. The main merits of our method are: (1) In order to obtain robust and accurate results under the situation of dramatic pose variation, we first evaluate the reliability of 2D tracker. Then after eliminating erroneous 2D correspondences, we refine the rigid motion parameters estimated between successive poses by performing a non-linear, batch estimator to compute the parameters of all poses in a clip of stereo sequences simultaneously. (2) Full automaticity is achieved by detecting and matching new features when there are not enough reliable 2D tracking results. Experiments show that this algorithm is accurate and robust, and help our system reach a satisfactory face modeling result.

Motion occurring during magnetic resonance imaging acquisition is a major factor of image quality degradation. Self-navigation can help reduce artefacts by estimating motion from the acquired data to enable motion correction. Popular self-navigation techniques rely on the availability of a fully-sampled motion-free reference to register the motion corrupted data with. In the proposed technique, rigid motion parameters are derived using the inherent correlation between radial segments in k-space. The registration is performed exclusively in k-space using the Phase Correlation Method, a popular registration technique in computer vision. Robust and accurate registration has been carried out from radial segments composed of as few as 32 profiles. Successful self-navigation has been performed on 2-D dynamic brain scans corrupted with continuous motion for six volunteers. Retrospective motion correction using the derived self-navigation parameters resulted in significant improvement of image quality compared to the conventional sliding window. This work also demonstrates the benefits of using a bit-reversed ordering scheme to limit undesirable effects specific to retrospective motion correction on radial trajectories. This method provides a fast and efficient mean of measuring rigid motion directly in k-space from dynamic radial data under continuous motion.

Purpose:Radiofrequency (RF) catheter ablation has transformed treatment for arrhythmias and has become first-line therapy for some tachycardias. The precise localization of the arrhythmogenic site and the positioning of the RF catheter over that site are problematic: they can impair the efficiency of the procedure and are time consuming (several hours). This study evaluates the feasibility of using only single plane C-arm images in order to estimate the 3D coordinates of RF catheter electrodes in a cardiac phase.Materials and methods:The method makes use of a priori 3D model of the RF mapping catheter assuming rigid body motion equations in order to estimate the 3D locations of the catheter tip-electrodes in single view C-arm fluoroscopy images. Validation is performed on both synthetic and clinical data using computer simulation models. The authors' monoplane reconstruction algorithm is applied to a 3D helix mimicking the shape of a catheter and undergoing solely rigid motion. Similarly, the authors test the feasibility of recovering nonrigid motion by applying their method on true 3D coordinates of 13 ventricular markers from a sheep’s ventricle.Results:The results of this study showed that the proposed monoplane algorithm recovers rigid motion adequately when using the spatial positions of a catheter in six consecutive C-arm image frames yielding maximum 3D root mean squares errors of 4.3 mm. On the other hand, the suggested algorithm did not recover nonrigid motion precisely as suggested by a maximum 3D root mean square value of 8 mm.Conclusion:Since RF catheter electrodes are rigid structures, the authors conclude that there is promise in recovering the 3D coordinates of the electrodes when making use of only single view images. Future work will involve adding nonrigid motion equations to their algorithm, which will then be applied to actual clinical data.

Tracking the trajectory of dribble directional shooting in the video image is beneficial to improve the basketball training skills. In this paper, the trajectory tracking method of dribble directional shooting based on the symmetrical algorithm is proposed. The symmetric differential algorithm is used to segment the video image of the dribble directional shooting. In the segmented image sequence, the range of motion of the dribble directional shooting is detected every three consecutive frames. Through the maximum variance ratio threshold method, the motion area of the dribble directional shot is divided. The initial motion coordinates and feature vectors of the dribble directional shooting are calculated, and the characteristic affine transformation equation of the shooting action pixel information is constructed. The Camshift method is used to mark the trajectory contour of the dribble directional shooting. According to the result of the contour extraction, the gray pixel value feature points of the video sequence are fitted, and the trajectory tracking model of the dribble directional shooting is established to accurately track the trajectory of the dribble directional shooting. It is verified that the trajectory of the dribble directional shooting tracked by the proposed method is consistent with the actual situation, and the tracking error average is about 1.9%, that is the reliable trajectory tracking method of dribble directional shooting.

This paper proposes an approach to estimate 3D rigid facial motions through a stereo image sequence. The approach uses a disparity space as the main space in order to represent all the 3D information. A robust algorithm based on the RANSAC approach is used to estimate the rigid motions through the image sequence. The disparity map is shown to be a robust feature against local motions of the surface and is therefore a very good alternative to the traditional use of the set of interest points.

Evaluation of functional magnetic resonance imaging (fMRI) as a reliable clinical imaging tool requires accurate assessment and correction of head motion artifacts. As the correction of bulk head motion is vital, the loss of signal strength from the confounding effect of head motion on spin magnetization may be an additional factor in activation analysis error. This study focuses on the evaluation and correction of the spin saturation artifact that occurs when parts of adjacent slices are selected due to changing head positions in single-shot multislice acquisitions. As a consequence of head movement, the acquired slices constituting a fMRI volume are no longer parallel to each other and the spin magnetization in fMRI voxels becomes dependent on head motion history. Motion corrections applying the same rigid motion estimates to all the slices in a volume may not be a reasonable approximation in cases where the magnitude of head motion exceeds a subvoxel range. For realistic ranges of motion in fMRI, an accurate estimate of rigid motion parameters for each echo planar imaging (EPI) slice is essential to correctly register voxel intensities. Previously we have implemented the map-slice-to-volume (MSV) motion correction method that maps each slice in a time series onto a reference anatomical volume, which proved to be effective in improving activation detection. To correctly evaluate the motion dependence of spin magnetization, each voxel is tracked with movement history that is available from MSV motion estimates. Relatively low in resolution, EPI voxels are composed of varying mixtures of white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) and variations in the tissue composition give rise to voxel intensities that are functions of tissue T1 properties. We have developed a weighted-average spin saturation (WASS) correction method that can handle full rigid motion and account for the melange of different brain tissue isochromats at each EPI voxel location. We evaluated the effect of spin saturation artifacts and the performance of the WASS correction using simulated fMRI time series synthesized with known true activation, motion, and the associated spin saturation artifact. Two different ranges of head rotations, [-5,5] and [-2,2] deg, were introduced and the effect of the spin saturation artifact was quantified to show 18% and 13% reduction in activation detection rate, respectively. Following the MSV motion and WASS correction, results indicate that WASS correction can improve activation detection by 17% relative to MSV only correction.

Despite the current availability in resource-rich regions of advanced technologies in scanning and 3-D imaging in current ophthalmology practice, world-wide screening tests for early detection and progression of glaucoma still consist of a variety of simple tools, including fundus image-based parameters such as CDR (cup to disc diameter ratio) and CAR (cup to disc area ratio), especially in resource -poor regions. Reliable automated computation of the relevant parameters from fundus image sequences requires robust non-rigid registration and segmentation techniques. Recent research work demonstrated that proper non-rigid registration of multi-view monocular fundus image sequences could result in acceptable segmentation of cup boundaries for automated computation of CAR and CDR. This research work introduces a composite diffeomorphic demons registration algorithm for segmentation of cup boundaries from a sequence of monocular images and compares the resulting CAR and CDR values with those computed manually by experts and from 3-D visualization of stereo pairs. Our preliminary results show that the automated computation of CDR and CAR from composite diffeomorphic segmentation of monocular image sequences yield values comparable with those from the other two techniques and thus may provide global healthcare with a cost-effective yet accurate tool for management of glaucoma in its early stage.

In this paper we propose a new method for variational segmentation of image sequences containing nonrigid, moving objects. The method is based on the Chan-Vese model augmented with a novel frame-to-frame interaction term, which allow us to update the segmentation result from one image frame to the next using the previous segmentation result as a shape prior. The interaction term is constructed to be pose-invariant and to allow moderate deformations in shape. It can handle the appearance of occlusions which otherwise can make segmentation fail. The performance of the model is illustrated with experiments on synthetic and real image sequences.

Image segmentation is an essential and important part in medical image processing. Conventional methods typically demand significant amounts of time and don¿t lend to a natural interaction scheme with 3D volume. In this paper we present a novel interactive method for fuzzy connectedness segmentation. In our approach, the user freely specifies a seed point 3D location at regions of interest directly over the 3D volume. Then the 3D seed points will be mapped into seed pixels in some 2D slices of image-sequences. After that, the proposed algorithm is used to segmentation of multiple objects from complex background and batch segmentation of image sequences. That can be achieved by seeds picked from object and background in 3D volume. The experimental results with a serial of abdominal CT images show that the proposed method can improve the accuracy of segmentation effectively with very little interaction.

The ModaVox program was developed to study acoustic propagation of vowel sounds in a 3D vocal tract by using the transmission line matrix (TLM) method. The name ModaVox stands for “Modelador da Voz” in Portuguese, or Voice Modeler in English. At the present moment, it is able to build and solve 3D TLM numerical models of the vocal tract. The meshes are constructed over the voxels (volumetric picture elements) in segmented medical image sequences. The segmentation of the images is performed via a neural network, island removal and some manual adjustments with the ModaVox’s toolboxes. An implementation of the TLM method allows for the simulation of the acoustic propagation of the input signal through the TLM mesh of the vocal tract model. ModaVox also generates tetrahedral and or surface triangle meshes.

Various approaches have been proposed for simultaneous optical flow estimation and segmentation in image sequences. In this study, the moving scene is decomposed into different regions with respect to their motion, by means of a pattern recognition scheme. The inputs of the proposed scheme are the feature vectors representing still image and motion information. Each class corresponds to a moving object. The classifier employed is the median radial basis function (MRBF) neural network. An error criterion function derived from the probability estimation theory and expressed as a function of the moving scene model is used as the cost function. Each basis function is activated by a certain image region. Marginal median and median of the absolute deviations from the median (MAD) estimators are employed for estimating the basis function parameters. The image regions associated with the basis functions are merged by the output units in order to identify moving objects.

One can segment a single digital image in grey level after having quantized it by a self-organising map. Because of the great computing time needed by the quantization process, one cannot quantize each image of a sequence when real time processing is at aim. We propose a method in which 1.- one has a set of precomputed maps among which is chosen the most appropriate one to segment the current image. Quantization errors determine the most appropriate map. 2.- one lets the current map continuously adapt itself to new images as long as it is the most appropriate. Time complexity and possibilities of parallelisation of the deduced algorithm are considered.