2D Image Sequences Research Articles

Estimation of 3D left ventricular deformation from echocardiography

The quantitative estimation of regional cardiac deformation from 3D image sequences has important clinical implications for the assessment of viability in the heart wall. Such estimates have so far been obtained almost exclusively from Magnetic Resonance (MR) images, specifically MR tagging. In this paper we describe a methodology for estimating cardiac deformations from 3D echocardiography (3DE). The images are segmented interactively and then initial correspondence is established using a shape-tracking approach. A dense motion field is then estimated using a transversely isotropic linear elastic model, which accounts for the fiber directions in the left ventricle. The dense motion field is in turn used to calculate the deformation of the heart wall in terms of strain in cardiac specific directions. The strains obtained using this approach in open-chest dogs before and after coronary occlusion, show good agreement with previously published results in the literature. They also exhibit a high correlation with strains produced in the same animals using implanted sonomicrometers. This proposed method provides quantitative regional 3D estimates of heart deformation from ultrasound images.

The Visual Looming Navigation Cue: A Unified Approach

Visual looming is related to an increasing projected size of an object on a viewer's retina as the relative distance between the viewer and the object decreases. It is an indication for threat that may be used along with visual fixation to accomplish navigation tasks. This paper defines visual looming as the time derivative of the relative distance (range) between the observer and the object divided by the relative distance itself. It introduces a unified approach to visual looming by showing how visual looming can be calculated from the relative temporal change in the following attributes of a 2D image sequence: (i) image area, (ii) image brightness, (iii) texture density in the image, and (iv) image blur. It is shown that a closed-form unified expression can be adopted in all these methods. Experimental results illustrate how the measured values of looming are related to the actual values. Finally, looming is used in the sense of a threat of collision, along with visual fixation, to navigate in an unknown environment.

Magnetic resonance hysterography and hysterosalpingography using hyperpolarized (3)He: demonstration of feasibility in an animal model.

Assessment of the uterine cavity and patency of the fallopian tubes remains a difficult goal with magnetic resonance imaging (MRI). The purpose of this paper is to describe the development of a new magnetic resonance hysterography (MR-HG) and hysterosalpingography (MR-HSG) technique employing hyperpolarized (3)He. Two-dimensional (2D) and 3D gradient-echo imaging sequences were developed and optimized using a phantom. An optimized sequence was then applied in swine cadavers. J. Magn. Reson. Imaging 2000;12:1009-1013.

Object-based coding of stereoscopic and 3D image sequences

The author begins by discussing what object based coding is and goes on to consider the structure of object based stereoscopic coders. Different techniques for object-based stereoscopic image sequence coding are reviewed. These techniques basically differ in the way they define models and estimate model parameters. We review the various models used for representing motion and structure. Then we review segmentation techniques, and discuss coding of object parameters and image synthesis.

Real-time obstacle avoidance using central flow divergence, and peripheral flow

The lure of using motion vision as a fundamental element in the perception of space drives this effort to use flow features as the sole cues for robot mobility. Real-time estimates of image flow and flow divergence provide the robot's sense of space. The robot steers down a conceptual corridor, comparing left and right peripheral flows. Large central flow divergence warns the robot of impending collisions at "dead ends". When this occurs, the robot turns around and resumes wandering. Behavior is generated by directly using flow-based information in the 2D image sequence. Active mechanical gaze stabilization simplifies the visual interpretation problems by reducing camera rotation. By combining corridor following and dead-end deflection, the robot has wandered around the lab at 30 cm/s for as long as 20 min without collision. The ability to support this behavior in real-time with current equipment promises expanded capabilities as computational power increases in the future.

Dense Non-Rigid Motion Estimation in Sequencesof Medical Images Using Differential Constraints

We describe a new method for computing the displacement vector field in time sequences of 2D or 3D images (4D data). The method is energy-minimizing on the space of correspondence functions; the energy is split into two terms, with one term matching differential singularities in the images, and the other constraining the regularity of the field. In order to reduce the computational time of the motion estimation, we use an adaptive image mesh, the resolution of which depends on the value of the gradient intensity. We solve numerically the minimization problem with the finite element method which gives a continuous approximation of the solution. We present experimental results on synthetic data and on medical images and we show how to use these results for analyzing cardiac deformations.

Analyzing facial expressions for virtual conferencing

The authors present a model-based algorithm that estimates 3D motion and facial expressions from 2D image sequences showing head and shoulder scenes typical of video telephone and teleconferencing applications.

Laser scanning confocal microscopy and factor analysis of biomedical image sequences (FAMIS) to detect and characterise HPV DNA sequences by FISH in HeLa cells

Visualisation and localisation of specific DNA sequences were performed by fluorescence in situ hybridisation (FISH), confocal laser scanning microscopy (CLSM), and factor analysis of biomedical image sequences (FAMIS). HeLa cells containing 10-50 copies per cell of human papillomavirus (HPV) DNA type 18 integrated in cellular DNA were used as a model. HPV-DNA was identified by DNA probes and DNA-DNA hybrids were revealed by alkaline phosphatase and Fast Red (FR) TR salt/naphtol-MX phosphate. Cell nuclei were counterstained with thiazole orange (TO). FAMIS summarises image sequences into a reduced number of images called factor images and curves called factors. Factor images correspond to spatial distributions of the different factors. Factors estimate different individual physical behaviours in the sequence (extinction velocity, spectral emission, depth emission profiles). We verified that HPV-DNA hybridisation signals are specific to the spectrum of FR, and distinguished between FR and TO. The latter result was found by taking into account differences in their extinction velocities. The focus of CLSM was improved on 3D image sequences, and the location of fluorescent signals inside the preparations was determined. Factor images showed that FR stained targets were located on different focal planes at the periphery of the nuclei.

Laser scanning confocal microscopy and factor analysis of biomedical image sequences (FAMIS) to detect and characterise HPV DNA sequences by FISH in HeLa cells

Visualisation and localisation of specific DNA sequences were performed by fluorescence in situ hybridisation (FISH), confocal laser scanning microscopy (CLSM), and factor analysis of biomedical image sequences (FAMIS). HeLa cells containing 10–50 copies per cell of human papillomavirus (HPV) DNA type 18 integrated in cellular DNA were used as a model. HPV-DNA was identified by DNA probes and DNA-DNA hybrids were revealed by alkaline phosphatase and Fast Red (FR) TR salt/naphtol-MX phosphate. Cell nuclei were counterstained with thiazole orange (TO). FAMIS summarises image sequences into a reduced number of images called factor images and curves called factors. Factor images correspond to spatial distributions of the different factors. Factors estimate different individual physical behaviours in the sequence (extinction velocity, spectral emission, depth emission profiles). We verified that HPV-DNA hybridisation signals are specific to the spectrum of FR, and distinguished between FR and TO. The latter result was found by taking into account differences in their extinction velocities. The focus of CLSM was improved on 3D image sequences, and the location of fluorescent signals inside the preparations was determined. Factor images showed that FR stained targets were located on different focal planes at the periphery of the nuclei. Cytometry 28:269–279, 1997. © 1997 Wiley-Liss, Inc.

Application of continuum theory and multi-grid methods to motion evaluation from 3D echocardiography

As the motion of the heart is a 3D phenomenon, its evaluation from sequences of 2D images causes a great loss of information on the motion itself. Our aim is therefore to process real 3D echocardiographic images and to carry out an automatic way of evaluating the movements of the cardiac structures. To estimate the optical flow, a mathematical model based on the continuum theory is used; echocardiographic images can indeed be considered a function of a conserved quantity (the acoustic impedance). Since we need to calculate the velocity vector for every point in the image and every image is built with more than 2 million voxels (128x128x128), we implement a multigrid relaxation method to accelerate the computation of an approximate solution otherwise too slow with a simple iterative solver. The experiments on simulated velocity fields have demonstrated an effective speed-up in the evaluation of motion, and the calculation on real echo images has given a realistic estimation of the 3D dynamics of the heart.

Coding of 3D moving medical data using a 3D warping technique

A novel motion compensation algorithm for the coding of 3D image sequences is presented. The temporal correlation existing between two successive in time 3D data sets is exploited using a 3D cube matching algorithm (CMA). In order to cope with more complex motions and to adapt compensation to the high activity regions, a 3D warping-based motion estimation technique is introduced. Simulation results are given for the 3D medical image sequence “Beating Heart”.

3D structure and motion estimation from 2D image sequences

Two novel algorithms are presented for depth estimation using point correspondences and the ground plane constraint. One is a direct non-iterative method, and the other a simple, well-behaved iterative technique where the choice of initial value is straightforward. The algorithms are capable of handling any number of points and frames as well as points which become occluded. Once the point depths are determined, motion parameters can be obtained by a linear least squares technique. Extensive test results are included which show that the proposed algorithms are robust to noise, and perform satisfactorily using real outdoor image sequences.

Book reviews

This monograph encompasses the extensive experience and expertise of its author, who has pioneered the technique of magnetic resonance imaging (MRI) of the breast using gadolinium enhancement, and has been eagerly awaited by those who practise in this field. Dr Heywang-Kobrunner's technique uses a 3D gradient echo fast imaging sequence with a slice thickness of no more than 5 mm.

Robot operation monitoring for collision avoidance by image sequence analysis

Safety is an important issue of industrial automation. In this study, a new approach to monitoring multiple operating robot manipulators for on-line collision avoidance by image sequence analysis is proposed. The features used are line segments and each two-dimensional (2D) projective robot shape forms a polygon. In the monitoring process, multiple image sequences about the operating robots are taken from various views using fixed video cameras. For each view, three image frames taken from three recent frames (current and previous two frames) are used to predict each robot shape location in the predicted space, i.e. in the next frame. Three-dimensional (3D) robot operation monitoring for collision avoidance is decomposed into some collision checkings performed in multiple 2D image sequences. In the predicted space, if two robot shapes overlap mutually or the minimal distance between two robot shapes is smaller than a thresholding value in all camera views, then the robots are forced to halt for safety; otherwise, they continue to operate. To speed up the monitoring system, line segment correspondence between two consecutive images from a camera view is performed via region matching. Region matching is formulated as a bipartite matching problem. To consider the effect of robot motion, a static distance thresholding T s and a dynamic distance thresholding T d are used to check if any two robot shapes are “safe” from a view. Some experimental results show the feasibility of the proposed approaches.