Showing papers presented at "British Machine Vision Conference in 1994"

Active shape models: evaluation of a multi-resolution method for improving image search

Abstract: We describe a multi-resolution technique for locating for variable structures in images. This is an extension of work on Active Shape Models (ASMs) - statistical models which iteratively deform to match image data. An ASM consists of a shape model controlling a set of landmark points, together with a statistical model of the grey-levels expected around each landmark. Both the shape model and the grey-level models are trained on sets of labelled example images. In order to apply a coarse-to-fine search strategy it is necessary to train a set of grey-level models for each landmark, one for every level of a multi-resolution image pyramid. During image search the model is started on the coarsest resolution image. As the search progresses it moves to finer and finer resolutions until no further improvement can be made. We describe an automatic technique for deciding when to :ss has converged. We demonstrate the ntitative experiments which show a sigi speed and quality of fit compared to previous methods.

An automatic face identification system using flexible appearance models

Abstract: We describe the use of flexible models for representing the shape and grey-level appearance of human faces. These models are controlled by a small number of parameters which can be used to code the overall appearance of a face for image compression and classification purposes. The model parameters control both inter-class and within-class variation. Discriminant analysis techniques are employed to enhance the effect of those parameters affecting inter-class variation, which are useful for classification. We have performed experiments on face coding and reconstruction and automatic face identification. Good recognition rates are obtained even when significant variation in lighting, expression and 3D viewpoint, is allowed. Human faces display significant variation in appearance due to changes in expression, 3D orientation, lighting conditions, hairstyles and so on. A successful automatic face identification system should be capable of suppressing the effect of these factors allowing any face image to be rendered expression-free with standardised 3D orientation and lighting. We describe how the variations in shape and grey-level appearance in face images can be modelled, and present results for a fully automatic face identification system which tolerates changes in expression, viewpoint and lighting.

Automatic landmark generation for Point Distribution Models

Abstract: Point Distribution Models (PDMs) are statistically derived flexible templates which are trained on sets of examples of the object(s) to be modelled. They require that each example is represented by a set of points (landmaiks) and that each landmark represents the same location on each of the examples. Generating the landmarks from 2D boundaries or 3D surfaces has previously been a manual process. Here, we describe a method for automatically generating PDMs from a training set of pixel- lated boundaries in 2D. The algorithm is a two-stage process in which a pair-wise corresponder is first used to establish an approximate set of landmarks on each of the example boundaries; in the second phase the landmarks are refined using an iterative non-linear optimisation scheme to generate a more compact PDM. We present results for two objects - the right hand and a chamber of the heart. The mo- dels generated using the automatically placed landmarks are shown to be better than those derived from landmarks located manually.

Statistical snakes: active region models

Abstract: This paper describes a new region-growing technique that uses a closed snake driven by a pressure force that is a function of the statistical characteristics of image data. This statistical snake expands until its elements encounter pixels that lie outside user-defined limits relative to a seed region; when these limits are violated the pressure force is reversed to make the model contract. Tension and stiffness forces keep the boundary of the region model smooth, and a repulsion force prevents self-intersection. Boundary elements can be added and removed in response to complexity changes, and the tension, stiffness and pressure parameters can be adjusted to preserve the energy balance of the changing model. Statistical snakes have been used to segment a variety of images including composite textures and NMR data volumes.

Euclidean structure from uncalibrated images

Abstract: A number of recent papers have demonstrated that camera “self-calibration” can be accomplished purely from image measurements, without requiring special calibration objects or known camera motion. We describe a method, based on self-calibration, for obtaining (scaled) Euclidean structure from multiple uncalibrated perspective images using only point matches between views. The method is in two stages. First, using an uncalibrated camera, structure is recovered up to an affine ambiguity from two views. Second, from one or more further views of this affine structure the camera intrinsic parameters are determined, and the structure ambiguity reduced to scaled Euclidean. The technique is independent of how the affine structure is obtained. We analyse its limitations and degeneracies. Results are given for images of real scenes. An application is described for active vision, where a Euclidean reconstruction is obtained during normal operation with an initially uncalibrated camera. Finally, it is demonstrated that Euclidean reconstruction can be obtained from a single perspective image of a repeated structure

A simple, intuitive camera calibration tool for natural images

Abstract: The paper reports an interactive tool for calibrating a camera, suitable for use in outdoor scenes. The motivation for the tool was the need to obtain an approximate calibration for images taken with no explicit calibration data. Such images are frequently presented to research laboratories, especially in surveillance applications, with a request to demonstrate algorithms. The method decomposes the calibration parameters into intuitively simple components, and relies on the operator interactively adjusting the parameter settings to achieve a visually acceptable agreement between a rectilinear calibration model and his own perception of the scene. Using the tool, we have been able to calibrate images of unknown scenes, taken with unknown cameras, in a matter of minutes. The standard of calibration has proved to be sufficient for model-based pose recovery and tracking of vehicles.

A non-linear generalisation of PDMs using polynomial regression

Abstract: We have previously described how to model shape variability by means of point distribution models (TDMs,) in which there is a linear relationship between a set of shape parameters and the positions of points on the shape. This linear formulation can fail for shapes which articulate or bend.' we show examples of such failure for both real and synthetic classes of shape. A new, more general formulation for PDMs, based on polynomial regression, is presented. The resulting Polynomial Regression PDMs (PRPDMsj perform well on the data for which the linear method failed.

Modelling object appearance using the grey-level surface

Abstract: We describe a new approach to modelling the appearance of structures in grey-level images. We assume that both the shape and grey-levels of the structures can vary from one image to another, and that a number of example images are available for training. A 2-D image can be thought of as a surface in 3 dimensions, with the third dimension being the grey-level intensity at each image point. We can represent the shape of this surface by planting landmark points across it. By examining the way such collections of points vary across different examples we can build a statistical model of the shape, which can be used to generate new examples, and to locate examples of the modelled structure in new images. We show examples of these composite appearance models and demonstrate their use in image interpretation.

Relief: pictorial and otherwise

Abstract: Surfaces play an important role in visual perception. They are perceived as «(perceptual) reliefs», that are surfaces in 2 + 1D perceptual space, that is the product space of the 2D visual field and the 1D «depth dimension». It is in many respects irrelevant whether the observer views a true 3D scene or a flat (2D) picture of a scene. In both cases, the percepts are reliefs in 2 + 1D perceptual space. In the latter case, one speaks of «pictorial relief». We discuss how perceptual reliefs can be measured and which aspects of these reliefs are especially robust against day-to-day intraobserver variations, changes of viewing conditions and interobserver differences. It turns out that only aspects of the partial depth order (based on depth precedence in infinitesimal regions) are stable. Thus, features of the relief are invariants of general «relief preserving transformations» that may actually scramble depth values at different locations. This is evident from the fact that human observers can only judge depth precedence with some degree of certainty for points that are on a single slope. We discuss the formal structure of these relief invariants. Important ones are the Morse critical points and the ridges and courses of the relief

A probabilistic fitness measure for deformable template models

Abstract: Methods for automatic image interpretation based on the use of deformable template models have proved very successful. Whatever deformable template scheme is used, one of the basic requirements is a method for assessing the likelihood that a particular model instance is the correct interpretation of a given image. We describe a Bayesian 'fitness' measure which combines the likelihood of the model shape with the evidential support in a principled way. Image search is carried out by minimising the fitness measure using multi-scale quasi-Newtonian optimisation. We have previously compared the performance of different fitness measures. Here we give results for the new method and show that, by making optimal use of the image evidence, it achieves more accurate interpretation than the best of the methods we have previously tested.

Hierarchical segmentation satisfying constraints

Abstract: A new hierarchical segmentation algorithm is described. Its computational complexity and memory requirements are detailed, showing it to be practicably applicable to images of useful size. A simple modification of the algorithm adapts it to produce hierarchical segmentations that satisfy a constraint set. Results are given showing that this adapted algorithm can be used as the basis of a semi-automatic object definition tool or as the interface between a low-level image description module and a high-level module coding for knowledge and expectation.

Fast Vehicle Localisation and Recognition Without Line Extraction and Matching

Abstract: A novel algorithm is presented in this paper for vehicle localisation and recognition under the ground-plane constraint. Unlike the vast majority of the existing modelbased object recognition schemes, the algorithm eliminates the need for explicit feature extraction and matching so that the computational cost is substantially lower. The algorithm is tested extensively with routine outdoor traffic images. Experimental results are included to illustrate the performance of the algorithm. The algorithm is developed for real-time implementation for applications in traffic scene analysis. It may readily be adapted to other industrial applications.

A Model Based Dual Active Contour

Abstract: Active contours are now established as a technique for extracting salient contours from an image. Unfortunately the original technique suffers from many problems. A novel model-based dual active contour, a method of integrating global shape information with two active contours, has been developed to overcome the primary problems; sensitivity to initialisation and undesirable attractions by insignificant localised or regionalised features. The model guides the technique to avoid insignificant minima and is relinquished when the energy minimum is sufficiently compatible. The technique then finally operates as a pair of conventional active contours, ensuring that only image information is extracted, consistent with the original technique.

Direct calibration and data consistency in 3-D laser scanning

Abstract: This paper addresses two aspects of triangulation-based range sensors using structured laser light: calibration and measurements consistency We present a direct calibration technique which does not require modelling any specific sensor component or phenomena, therefore is not limited in accuracy by the inability to model error sources We also sketch some consistency tests based on two-camera geometry which make it possible to acquire satisfactory range images of highly reflective surfaces with holes Experimental results indicating the validity of the methods are reported

Combining point distribution models with shape models based on finite element analysis

Abstract: This paper describes a method of combining two approaches to modelling flexible objects. Modal Analysis using Finite Element Methods (FEMs) generates a set of vibrational modes for a single shape. Point Distribution Models (PDMs) generate a statistical model of shape and shape variation from a set of example shapes. A new approach is described which generates vibrational modes when few example shapes are available and changes smoothly to using more statistical modes of variation when a large data set is presented. Results are given for both synthetic and real examples. Experiments using the models for image search show that the combined version performs better than either the PDM or FEM models alone.

Grouping Curved Lines.

Abstract: This paper examines the problem of automatically grouping image curves. In contrast, most previous work has been restricted to points and straight lines. Some of the computational aspects of the groupings of continuation, parallelism, and proximity are analysed, and the issues of neighbourhoods, combinatorix, and multiple scales are discussed.

Slime - a new deformable surface

Abstract: Deformable surfaces have many applications in surface reconstruction, tracking and segmentation of range or volumetric data. Many existing deformable surfaces connect control points in a predefined and inflexible way. This means that the surface topology is fixed in advance, and also imposes severe limitations on how a surface can be described. For example a rectangular grid of control points cannot be evenly distributed over a sphere, and singularities occur at the poles. In this paper we introduce a new (G continuous) deformable surface. In contrast to other methods this method can represent a surface of arbitrary topology, and do so in an efficient way. The method is based on a generalization of biquadratic B-splines, and has a comparable computational cost to methods based on traditional tensor product B-splines.

Viewpoint-invariant representation of generalized cylinders using the symmetry set

Abstract: We demonstrate that viewpoint-invariant representations can be obtained from images for a useful class of 3D smooth object. The class of surfaces are those generated as the envelope of a sphere of varying radius swept along an axis. This class includes canal surfaces and surfaces of revolution. The representations are computed, using only image information, from the symmetry set of the object's outline. They are viewpoint-invariant under weak-perspective imaging, and quasi-invariant to an excellent approximation under perspective imaging. To this approximation, the planar axis of a canal surface is recovered up to an affine ambiguity from perspective images. Examples are given of the representations obtained from real images, which demonstrate stability and object discrimination, for both canal surfaces and surfaces of revolution. Finally, the representations are used as the basis for a model-based object recognition system

Intensity-Based Object Extraction from 3D Medical Images Including a Correction of Partial Volume Errors.

Abstract: Feature extraction by applying a threshold window to image intensity values is a simple and common image processing technique. We consider the case of 3D images where intensity based feature extraction is used to determine the volume of objects of interest. We show that the accuracy of volume determination is limited by partial volume (PV) effects. We outline a new method for correcting for PV effects based on object geometry and object intensity. Although this PV correction has been developed with respect to a specific application in magnetic resonance imaging, it is applicable to volume extraction by thresholding in any image and can possibly be extended to other intensity-based extraction techniques.

Illumination invariant colour recognition

Abstract: The article describes a colour based recognition system with three novel features. Firstly, the proposed system can operate in environments where spectral characteristics of illumination change in both space and time. Secondly, benefits in terms of speed and quality of output are gained by focusing processing to areas of salient colour. Finally, an automatic model acquisition procedure allows rapid creation of the model database.

Systematic Methods for Multivariate Data Visualization and Numerical Assessment of Class Separability and Overlap in Automated Visual Industrial Quality Control

Abstract: The focus of this work is on systematic methods for the visualization and quality assessment with regard to classificatio n of multivariate data sets. Our novel methods and criteria give in visual and numerical form rapid insight in the principal data distribution, the degree of compactness and overlap of class regions and class separability, as well as information to identify outliers in the data set and trace them back to data acquisition. Assessment by visualization and numerical criteria can be exploited for interactive or automatic optimization of feature generation and selection/extraction in pattern recognition problems. Further, we provide a novel criterion to assess the credibility and reliability of the visualization obtained from high dimensional data projection. Our methods will be demonstrated using data from visual industrial quality control and mechatronic applications.

Skewed symmetry detection through local skewed symmetries

Abstract: We explore how global symmetry can be detected prior to segmentation and under noise and occlusion. The definition of local symmetries is extended to affine geometries by considering the tangents and curvatures of local structures, and a quantitative measure of local symmetry known as symmetricity is introduced, which is based on Mahalanobis distances from the tangent-curvature states of local structures to the local skewed symmetry state-subspace. These symmetricity values, together with the associated local axes of symmetry, are spatially related in the local skewed symmetry field (LSSF). In the implementation, a fast, local symmetry detection algorithm allows initial hypotheses for the symmetry axis to be generated through the use of a modified Hough transform. This is then improved upon by maximising a global symmetry measure based on accumulated local support in the LSSF — a straight active contour model is used for this purpose. This produces useful estimates for the axis of symmetry and the angle of skew in the presence of contour fragmentation, artifacts and occlusion.

Image registration using multi-scale texture moments

Abstract: In this paper we propose a novel, efficient and geometrically intuitive method to compute the four components of an affine transformation from the change in simple statistics of images of texture. In particular, we show how the changes in second circular moments of edge orientation are directly related to the rotation (curl), scale (divergence) and deformation components of an affine transformation, and how these components can be computed from multi-scale texture moments. A simple implementation is described which does not require point, edge or contour correspondences to be established. It is tested on repetitive and non-repetitive visual textures which are neither isotropic nor homogeneous. The theoretical accuracy and the noise sensitivity of this method are compared with other linear moment and circular moment methods.

Computing Covering Polyhedra of Non-Convex Objects

Abstract: We present an algorithm to build covering polyhedra for digital 3D objects, by iteratively filling local concavities. The resulting covering polyhedron is convex and is a good approximation of the convex hull of the object. The algorithm uses 3x3x3 operators and requires a few minutes for a 128 x 128 x 128 image, when implemented on a sequential computer. Once the covering polyhedron has been obtained, the object concavities can be identified by subtracting the object from the polyhedron and suitably post-processing the set difference. Features characterising the concavities can then be extracted and used as a tool for quantitative shape analysis.

Detecting floor anomalies

Abstract: When a robot moves about a 2D world such as a planar surface, it is important that obstacles to the robot's motions be detected. This classical problem of "obstacle detection" has proven to be difficult. Many researchers have formulated this problem as being the process of determining where a robot cannot move due to the presence of obstacles. An alternative approach presented here is to determine where an robot can go by identifying floor regions for which the planar floor assumption can be verified. A stereo vision system is developed for Floor Anomaly Detection (FAD), and its relationship to existing stereo obstacle detection algorithms is described.

Application of an Associative Memory to the Analysis of Document Fax Images.

Abstract: An essential part of image analysis is the identification of objects within the image. This paper describes the results of applying a binary associative memory neural network (ADAM), to the complex task of identifying shapes in document images. The associative memory is used to implement the generalised Hough Transform, exploiting the fast look-up ability of the associative memory to give a high-speed image analysis tool.

Dynamic fixation of a moving surface using log polar sampling

Abstract: We describe the development and testing of a first-order motion estimation algorithm which maintains accurate fixation of features on surfaces undergoing three-dimensional motion, and determines the local affine motion parallax. The accuracy of the first-order flow estimation is much improved by the use of log-polar sampling. We investigate the contribution of fixation to this accuracy using synthetic flow, and demonstrate the performance on affine tracking in real image sequences.

An evolutionary algorithm for physical motion analysis

Abstract: This paper presents an approach to motion understanding, through identification of physical pa- rameters from image sequences. It is based on a family of particle-base d physical models where deformable objects are represented as sets of weighted particles and their interactions. The inter- action model presented derives from an energy potential, using dual bonds (extension springs) and ternary bonds (torsional springs). An original dynamical motion analysis algorithm is described, which extracts physical animation parameters (springs lengths, angles, stiffness...) through the processing of an image sequence. Ge- netic techniques are employed to perform the fitting of parameters in an analysis-by-synthesis scheme. Experimental test results on synthetic sequences are reported.

Using Voxel Similarity as a Measure of Medical Image Registration

Abstract: In this paper we present our work on using intensity feature spaces to study the relationship between voxel values in registered and unregistered medical images. By taking a simple image model we predict structures that we might expect to find in an intensity feature space produced from different modality images of the same scene. We show how this structure will be modified by image noise, misregistration and differing point spread functions of the two modalities. We show examples of such structure in feature spaces created from clinically acquired Magnetic Resonance (MR) and Positron Emission Tomography (PET) image data. We show how two simple measures of voxel similarity based on these feature space observations, a modified variance of intensity ratio and the 3rd order moment of the feature space histogram, can be used to quantify image misregistration. The 3rd order moment measure is then used with a genetic optimisation algorithm to automatically register pre and post Gadolinium injection MR images.