Showing papers presented at "British Machine Vision Conference in 1995"

Learning the distribution of object trajectories for event recognition

Abstract: The advent in recent years of robust, real-time, model-based tracking techniques for rigid and non-rigid moving objects has made automated surveillance and event recognition a possibility. A statistically based model of object trajectories is presented which is learnt from the observation of long image sequences. Trajectory data is supplied by a tracker using Active Shape Models, from which a model of the distribution of typical trajectories is learnt. Experimental results are included to show the generation of the model for trajectories within a pedestrian scene. We indicate how the resulting model can be used for the identification of atypical events.

Image difference threshold strategies and shadow detection

Abstract: The paper considers two problems associated with the detection and classification of motion in image sequences obtained from a static camera. Motion is detected by differencing a reference and the "current" image frame, and therefore requires a suitable reference image and the selection of an appropriate detection threshold. Several threshold selection methods are investigated, and an algorithm based on hysteresis thresholding is shown to give acceptably good results over a number of test image sets. The second part of the paper examines the problem of detecting shadow regions within the image which are associated with the object motion. This is based on the notion of a shadow as a semi-transpare nt region in the image which retains a (reduced contrast) representation of the underlying surface pattern, texture or grey value. The method uses a region growing algorithm which uses a growing criterion based on a fixed attenuation of the photometric gain over the shadow region, in comparison to the reference image.

A buyer's guide to conic fitting

Abstract: In this paper we evaluate several methods of fitting data to conic sections. Conic fitting is a commonly required task in machine vision, but many algorithms perform badly on incomplete or noisy data. We evaluate several algorithms under various noise and degeneracy conditions, identify the key parameters which affect sensitivity, and present the results of comparative experiments which emphasize the algorithms' behaviours under common examples of degenerate data. In addition, complexity analyses in terms of flop counts are provided in order to further inform the choice of algorithm for a specific application.

A comparison of four algorithms for estimating 3-D rigid transformations

Abstract: A common need in machine vision is to compute the 3-D rigid transformation that exists between two sets of points for which corresponding pairs have been determined. In this paper a comparative analysis of four popular and efficient algorithms is given. Each computes the translational and rotational components of the transform in closed-form as the solution to a least squares formulation of the problem. They differ in terms of the representation of the transform and the method of solution, using respectively: singular value decomposition of a matrix, orthonormal matrices, unit quaternions and dual quaternions. This comparison presents results of several experiments designed to determine the (1) accuracy in the presence of noise, (2) stability with respect to degenerate data sets, and (3) relative computation time of each approach.

A generic deformable model for vehicle recognition

Abstract: This paper reports the development of a highly parameterised 3-D model able to adopt the shapes of a wide variety of different classes of vehicles (cars, vans, buses, etc), and its subsequent specialisation to a generic car class which accounts for most commonly encountered types of car (includng saloon, hatchback and estate cars) An interactive tool has been developed to obtain sample data for vehicles from video images A PCA description of the manually sampled data provides a deformable model in which a single instance is described as a 6 parameter vector Both the pose and the structure of a car can be recovered by fitting the PCA model to an image The recovered description is sufficiently accurate to discriminate between vehicle sub-classes

Active shape models and the shape approximation problem

Abstract: Active Shape Models (ASM) use an iterative algorithm to match statistically defined models of known but variable objects to instances in images. Each iteration of ASM search involves two steps: image data interrogation and shape approximation. Here we consider the shape approximation step in detail. We present a new method of shape approximation which uses directional constraints. We show how the error term for the shape approximation problem can be extended to cope with directional constraints, and present iterative solutions to the 2D and 3D problems. We also present an efficient algorithm for the 2D problem in which a modification of the error term permits a closed-form approximate solution which can be used to produce starting estimates for the iterative solution.

Robust recognition of scaled shapes using pairwise geometric histograms

Abstract: The recognition of shapes in images using Pairwise Geometric Histograms has previously been confined to fixed scale shape. Although the geometric representation used in this algorithm is not scale invariant, the stable behaviour of the similarity metric as shapes are scaled enables the method to be extended to the recognition of shapes over a range of scale. In this paper the necessary additions to the existing algorithm are described and the technique is demonstrated on real image data. Hypotheses generated by matching scene shape data to models have previously been resolved using the generalised Hough transform. The robustness of this method can be attributed to its approximation of maximum likelihood statistics. To further improve the robustness of the recognition algorithm and to improve the accuracy to which an objects location, orientation and scale can be determined the generalised Hough transform has been replaced by the probabilistic Hough transform.

An adaptive eigenshape model

Abstract: There has been a great deal of recent interest in statistical models of 2D landmark data for generating compact deformable models of a given object. This paper extends this work to a class of parametrised shapes where there are no landmarks available. A rigorous statistical framework for the eigenshape model is introduced, which is an extension to the conventional Linear Point Distribution Model. One of the problems associated with landmark free methods is that a large degree of variability in any shape descriptor may be due to the choice of parametrisation. An automated training method is described which utilises an iterative feedback method to overcome this problem. The result is an automatically generated compact linear shape model. The model has been successfully applied to a problem of tracking the outline of a walking pedestrian in real time.

Generating spatiotemporal models from examples

Abstract: Recent work on contour tracking has shown the benefits of using learned dynamic models for robust curve tracking. In this paper, we extend this work by using a physically based framework to learn physically plausible, constrained dynamic models of deforming objects. Traditional physically based vibration modes have been shown to provide a useful mechanism for describing non-rigid motions of articulated and deformable objects. The standard approach relies on assumptions being made about the elastic properties of an object to generate a compact set of real, orthogonal shape parameters which can then be used for tracking and data approximation. We present a method for automatically generating an improved physically based model using a training set of examples of the object deforming, tuning the elastic properties of the object to reflect how the object actually deforms. The resulting model provides a low dimensional shape description that allows accurate temporal extrapolation at low computational cost based on the training motions. Results are shown in which the method is applied to an automatically acquired training set of the outline of a walking pedestrian.

Automated 3D registration of truncated MR and CT images of the head

Abstract: This paper presents a fully automated technique for the rigid body registration of clinically acquired MR and CT images of the head. We describe our multiresolution approach to the optimisation of voxel similarity measures and evaluate the performance of a number of measures when presented with clinical images with a small overlapping volume. Results are compared with those derived from manual registration by identification of corresponding point landmarks. We show that by limiting the measures to intra-cranial regions of the images, not containing deformable skin surface features, a greater accuracy may be provided for certain types of truncated image.

Non-linear point distribution modelling using a multi-layer perceptron

Abstract: Objects of the same class sometimes exhibit variation in shape. This shape variation has previously been modelled by means of point distribution models (PDMs) in which there is a linear relationship between a set of shape parameters and the positions of points on the shape. A polynomial regression generalization of PDMs, which succeeds in capturing certain forms of non-linear shape variability, has also been described. Here we present a new form of PDM, which uses a multi-layer perceptron to carry out non-linear principal component analysis. We compare the performance of the new model with that of the existing models on two classes of variable shape: one exhibits bending, and the other exhibits complete rotation. The linear PDM fails on both classes of shape; the polynomial regression model succeeds for the first class of shapes but fails for the second; the new multi-layer perceptron model performs well for both classes of shape. The new model is the most general formulation for PDMs which has been proposed to date.

Tracking and measuring drivers eyes

Abstract: We describe a computer vision system for tracking the eyes of a car driver in order to measure the eyelid separation. This measure is used as part of a larger system designed to detect when a car driver is becoming drowsy. The system runs unattended in a car on modest hardware, does not interfere with the driver's normal driving actions, and requires no co-operation from the driver.

Statistical grey-level models for object location and identification

Abstract: This paper presents a new method for modelling and locating objects in images for applications such as Printed Circuit Board (PCB) inspection. Objects of interest are assumed to exhibit little variation in size or shape from one example to the next, but may vary considerably in grey-level appearance. Simple correlation based approaches perform poorly on such examples. To deal with variation we build statistical models of the grey levels across the structure in a set of training examples. A multi-resolution search technique is used to locate the best match to the model in an area of a new image to sub-pixel accuracy. A fit measure with predictable statistical properties can then be used to determine the probability that best match is a valid example of the model. We describe a ‘bootstrap’ approach to training and a method of automatically refining the final model to improve its performance. We demonstrate the method on PCB inspection, showing the approach is robust enough for use in a real production environment.

Robust matching by partial correlation

Abstract: Stereo matching by correlation near occlusions is a very challenging problem. When a partial occlusion occurs, most of the standard methods fail to produce acceptable results. This is because the techniques used do not take into account the presence of the occluding region. We propose a robust technique which we call partial correlation. This technique makes use of the least median square method which has recently been used for vision problems such as robust surface reconstruction. It performs better than standard methods near occlusions. It works by first of all disambiguating between the occluding region and the object region in the template and in the candidate window. A binary weighted correlation is then performed on the object regions. We present a comparative study between our approach and two other techniques. Experiment results validate our approach.

Statistical background modelling for tracking with a virtual camera

Abstract: A method of robust feature-detection is proposed for visual tracking with a pan-tilt head. Even with good foreground models, the tracking process is liable to be disrupted by strong features in the background. Previous researchers have shown that the disruption can be somewhat suppressed by the use of image-subtraction. Building on this idea, a more powerful statistical model of background intensity is proposed in which a Gaussian mixture distribution is fitted to each of the pixels on a "virtual" image plane. A fitting algorithm of the "Expectation-Maximisation" type proves to be particularly effective here. Practical tests with contour tracking show marked improvement over image subtraction methods. Since the burden of computation is off-line, the online tracking process can run in real-time, at video field-rate.

Tracking and recognising hand gestures using statistical shape models

Abstract: Hand gesture recognition from video images is of considerable interest as a means of providing simple and intuitive man-machine interfaces. Possible applications range from replacing the mouse as a pointing device to virtual reality and communication with the deaf. We describe an approach to tracking a hand in an image sequence and recognising, in each video frame, which of five gestures it has adopted. A statistically based Point Distribution Model (PDM) is used to provide a compact parameterised description of the shape of the hand for any of the gestures or the transitions between them. The values of the resulting shape parameters are used in a statistical classifier to identify gestures. The model can be used as a deformable template to track a hand through a video sequence but this proves unreliable. We describe how a set of models, one for each of the five gestures, can be used for tracking with the appropriate model selected automatically. We shown that this results in reliable tracking and gesture recognition for two 'unseen' video sequences in which all the gestures are used.

Detection and tracking of independent motion

Abstract: We describe an efficient method of using a translating camera to detect and track independently translating objects and assess the likelihood of a collision. By analysing the underlying geometry, it is shown that the tracking is reduced to two independent linear searches for a single feature in the image plane. Results are presented for both an off-line and a real time implementation using no special hardware. The method is completely automatic and shown to be accurate and robust.

Texture anisotropy, symmetry, regularity: recovering structure and orientation from interaction maps

Abstract: We discuss a novel method for recovering fundamental, perceptually motivated structural features of a texture pattern: anisotropy, symmetry, and regularity. The method is based on extended spatial grey-level difference statistics which describe pairwise pixel interactions and yield an interaction map used to assess the overall two-dimensional structure of interactions and extract the significant short- and long-range interactions (intersample spacings). The new approach extends, in digital images, the notion of greylevel difference to arbitrary spacing vectors (i.e. any angle at any displacement). This provides the necessary background for precise anisotropy (or directionality) and symmetry analysis. Experimental results are shown with a set of Brodatz images that range from highly regular to patterns with weak regularity or anisotropy. A few especially interesting examples of recovering hardly visible structural features are given. Finally, the approach is applied to rotation-invariant texture classification.

Active shape models and the shape approximation problem.

Abstract: The Active Shape Model(ASM) is an iterative algorithm for image interpretation based upon a Point Distribution Model. Each iteration of the ASM has two steps: Image data interrogation followed by shape approximation. Here we consider the shape approximation step in detail. We present a new method of shape approximation which uses directional constraints. We show how the error term for the shape approximation problem can be extended to cope with directional constraints and present iterative solutions to the 2D and 3D problems. We also show how the error term can be modified to allow a closed solution in the 2D case.

Equal-distance sampling of superellipse models

Abstract: Superellipses are parametric models that can be used for representing two dimensional object parts or aspects of 3-D parts. Previously little care was given to obtaining a precise sampling of the contour of these models. Equal-distance sampling of superellipse model contours is however important for rendering and in cases in which a cost function needs to be estimated for data fitting or parameter estimation, such as in model-based optimisation. In this paper we present a new parametric method for achieving equal-distance sampling of superellipse model contours that properly combines two simple first order models of the sampled points distance function. We also show how to extend the method to deformable superellipses and superquadrics.

Contour fitting using an adaptive spline model

Abstract: This paper presents a new segmentation algorithm by fitting active contour models (or snakes) to objects using adaptive splines. The adaptive spline model describes the contour of an object by a set of piecewisely interpolating C polynomial spline patches which are locally controlled. Thus the resulting description of the object contour is continuous and smooth. Polynomial splines provide a fast and efficient way for interpolating the object contour and allow us to compute its internal energy due to bending and elasticity deformations analytically. The adaptive spline model can be represented by its spline control points. The accuracy of the model is gradually increased during the segmentation process by inserting new control points. For estimating the optimal position of the control points, two different relaxation techniques based on Markov-Random-Fields (MRFs) have been combined and evaluated: Simulated Annealing (SA), which is a stochastic relaxation technique, and Iterated Conditional Modes (ICM), which is a probabilistic relaxation technique. We have studied convergence behavior and performance on artificial and medical images. The results show that the combination of both relaxation techniques provides very robust and initialization independent segmentation results.

Active contour models for shape description using multiscale differential invariants

Abstract: Classic curvature-minimizing active contour models are often incapable of extracting complex shapes with points of high curvature. This paper presents a new active contour model which overcomes this problem and which can be applied to image segmentation as well as shape description in order to allow for quantitative and qualitative studies of shape measurements at multiple scales. Multiscale differential operators, which are invariant to linear intensity transformations such as contrast or brightness adjustments and independent of coordinate transformations, are integrated into the model's spline energy functional. Whereas the image intensity gradient attracts the spline contour to image features, the isophote curvature of the image intensity function is used for matching the contour curvature. This novel curvature matching approach appears to be very useful for the extraction of very complex and strongly curved objects such as brain contours, results of which will be presented in this paper.

Automated pivot location for the Cartesian-Polar hybrid point distribution model

Abstract: The Point Distribution Model (PDM) has already proved useful for many tasks involving the location or tracking of deformable objects. A principal limitation is that non-linear variations must be approximated by combining linear variations, which sometimes results in a non-optimal model producing implausible object shapes. The Cartesian-Polar Hybrid PDM helps to overcome this limitation; selective use of polar geometry allows bending or pivotal deformation to be modelled more accurately; model components which exhibit no such trend remain in the Cartesian domain. Use of the Hybrid PDM currently requires the identification of pivot points by hand. In this paper we present a method for automatically identifying rigid model parts and pivot points from the training data. Experimental results are given for real data from human hands and for synthetic data from a simple jointed object.

Multi-variate cross-correlation and image matching

Abstract: This paper introduces the use of a multi-variate correlation function for region-based image matching and extends it to a modified crosscorrelation function that works well when matching image areas are required have the same intensity contrast. It also shows that the multivariate case is a straightforward generalisation of the monochrome image case. Experiments with both MRI and RGB colour imagery are shown, along with comparisons with the Euclidean, Manhatten and Loo matching metrics.

Convex hulls, occluding contours, aspect graphs and the Hough transform

Abstract: The Hough transform is a standard technique for finding features such as lines in images. Typically, edgels or other features are mapped into a partitioned parameter or Hough space as individual votes. The target image features are detected as peaks in the Hough space. In this paper we consider not just the peaks but the mapping of the entire shape boundary from image space to the Hough parameter space. We analyse this mapping and illustrate correspondences between features in Hough space and image space. Using this knowledge we present an algorithm to construct convex hulls of arbitrary 2D shapes with smooth and polygonal boundaries as well as isolated point sets. We also demonstrate its extension to the 3D case. We then show how this mapping changes as we move the origin in image space. The origin can be considered as a vantage point from which to view the object, and the occluding contour can be extracted easily from Hough space as those points where R = 0. We demonstrate the potential for tracking of transitions in the mapping to be used to construct an aspect graph of arbitrary 2D and 3D shapes.

Towards an automatic human face localization system

Abstract: This paper describes a method to detect and locate human faces in an image given no prior information about the size, orientation, and viewpoint of the faces in the image. This method uses a family of Gaussian derivative filters to search and extract human facial features from the image and then group them together into a set of partial faces using their geometric relationship. A belief network is then constructed for each possible face candidate and the belief values updated by evidences propagating through the network. Different instances of detected faces are then compared using their belief values and improbable face candidates discarded. The algorithm is tested on different instances of faces with varying sizes, orientation and viewpoint and the results indicate a 91% success rate in detection under viewpoint variation.

Self-alignment of a binocular robot

Abstract: We describe a novel method by which a four-axis binocular head platform can autonomously align its cameras so that their optic axes are parallel to each other and to the forward direction of the robot. The method uses controlled pans and elevations of the robot while viewing an unstructured scene to determine lines on the plane at infinity, whose intersection we prove to be the forward direction of the robot. The alignment is completed by fixating the projections of this point in both cameras. We summarize the underlying theory, and present results from a fully autonomous implementation of the algorithm.

Robust computation of optical flow

Abstract: This paper presents an algorithm to compute optical flow accurately at motion discontinuities and occlusion regions based on a robust estimator (the Least-Median-Squares estimator). The motion constraint equation and the 2-D affine motion model are used to compute the optical flow in a local neighbourhood. The use of the least-median-squares robust estimator enables points where optical flow cannot be computed to be rejected as outliers rather than assigning erroneous flow to such points. In addition, the use of an overlapping neighbourhood strategy eliminates the blockeffects that are commonly faced in local differential methods for computing optical flow. The algorithm is also able to deal with cases of the local neighbourhood straddling regions of three motions. Results for both synthetic and real image sequences are presented.

Medical image registration incorporating deformations

Abstract: Multiple sources of 3D medical image data can be used to construct detailed patient representations. Typically registration is achieved assuming the validity of rigid body transformation. In many applications, and in particular when updating representations used for guidance during surgery and therapeutic interventions, this assumption is inappropriate. In this paper we describe a general method for 3D deformation, show how registration can incorporate a composite of rigid body and deformation components and illustrate this methodology on 3 example sets of images. The first is a repeated 3D MR scan of the abdomen of a volunteer who purposely changed position between scans; the second is an MR and CT scan of the head and neck, in which the patient was in a different position for the two scans; and the third is a set of MR and CT images of the head taken before and after epilepsy surgery. Non rigid deformation and composite warping showed significant improvement in registration accuracy in each case.