This paper reports novel algorithms for the efficient localization and recognition of traffic in traffic scenes. The algorithms eliminate the need for explicit symbolic feature extraction and matching. The pose and class of an object is determined by a form of voting and one-dimensional (1-D) correlations based directly on image gradient data, which can be computed "on the fly." The algorithms are therefore very well suited to real-time implementation. The algorithms make use of two a priori sources of knowledge about the scene and the objects expected: (1) the ground-plane constraint and (2) the fact that the overall shape of road vehicles is strongly rectilinear. Additional efficiency is derived from making the weak perspective assumption. These assumptions are valid in the road traffic application domain. The algorithms are demonstrated and tested using routine outdoor traffic images. Success with a variety of vehicles in several traffic scenes demonstrates the efficiency and robustness of context-based image understanding in road traffic scene analysis. The limitations of the algorithms are also addressed in the paper.

This paper presents a detailed comparative study of 4 rotation invariant texture analysis methods. Human subjects are included as a benchmark for the computational methods. Experiments are conducted on two databases of 450 and 1320 images taken from 10 and 44 Brodatz texture classes respectively. Classification and content based retrieval experiments are used in the comparison, which includes the effect of Gaussian noise on each method (including the human subjects). The methods tested are: the multichannel Gabor filtering method; an edge attribute processing method; the circular simultaneous autoregressive (CSAR) model method and a method based on hidden Markov models with wavelet decomposition. The test conditions in the study are kept both general and constant so direct comparisons of the results obtained can be made. 1. Introduction Texture analysis is important for many applications such as content based image retrieval and scene analysis. Research on the subject has ...

Many techniques have been reported for handwriting-based writer identification. Most such techniques assume that the written text is fixed (e.g., in signature verification). In this paper we attempt to eliminate this assumption by presenting a novel algorithm for automatic text-independent writer identification from non-uniformly skewed handwriting images. Given that the handwriting of different people is often visually distinctive, we take a global approach based on texture analysis, where each writers’ handwriting is regarded as a different texture. In principle this allows us to apply any standard texture recognition algorithm for the task (e.g., the multi-channel Gabor filtering technique). Results of 96.0% accuracy on the classification of 150 test documents from 10 writers are very promising. The method is shown to be robust to noise and contents.

The paper presents an agent based surveillance system for use in monitoring scenes involving both pedestrians and vehicles. The system supplies textual descriptions for the dynamic activity occurring in the 3D world. These are derived by means of dynamic and probabilistic inference based on geometric information provided by a vision system that tracks vehicles and pedestrians. The symbolic scene annotation is given at two major levels of description: the object level and the inter-object level. At object level, each tracked pedestrian or vehicle is assigned a behaviour agent which uses a Bayesian network to infer the fundamental features of the objects' trajectory, and continuously updates its textual description. The inter-object interaction level is interpreted by a situation agent which is created dynamically when two objects are in close proximity. In the work included here the situation agent can describe a two-object interaction in terms of basic textual annotations, to summarise the dynamics of the local action.

A method of detecting and tracking objects using colour is presented. We track objects using a simple colour histogram based technique, which is fast and more robust to occlusion and camera motion than simple "blob" based techniques.

Many techniques have been reported for handwriting-based writer identification. Most techniques assume that the written text is fixed (e.g., in signature verification). In this paper we attempt to eliminate this assumption by presenting a novel algorithm for automatic text-independent writer identification. Given that the handwriting of different people can often be visually distinctive, we take a global approach based on texture analysis, where each writer's handwriting is regarded as a different texture. In principle this allows us to apply any standard texture recognition algorithm for the task (e.g., the multichannel Gabor filtering technique). Results of 95.0% accuracy on the classification of 300 test documents front 20 writers are very promising. The method is shown to be robust to noise and contents.

A novel algorithm is presented for determining the orientation of road vehicles in traffic scenes using video images. The algorithm requires no specific 3-D vehicle models and only uses local image gradient values. It may easily be implemented in real-time. Experimental results with a variety of vehicles in routine traffic scenes are included to demonstrate the effectiveness of the algorithm.

This paper reports novel algorithms for the efficient localisation and recognition of vehicles in traffic scenes, which eliminate the need for explicit symbolic feature extraction and matching. The algorithms make use of two a priori sources of knowledge about the scene and the objects: (i) the ground-plane constraint, and (ii) the fact that road vehicles are strongly rectilineal: The algorithms are demonstrated and tested using routine outdoor traffic images. Success with a variety of vehicles demonstrates the efficiency and robustness of context-based computer vision in road traffic scenes. The limitations of the algorithms are also addressed in the paper.

This paper uses Petri nets as a representation and analysis framework for automated assembly. It shows how the hierarchical nature of assembly is captured by Petri nets in representing assembly plans and lower level control plans. The tasks in a control plan correspond to assembly robot operations and can be viewed as a lower level representation of tasks at the assembly plan level. Petri nets provide a single unified framework for representing the hierarchical nature of assembly. This paper presents an approach to deriving performance and throughput measures of assembly on the basis of critical path analysis of the Petri net representation of plans.

This paper presents a boundary graph representation to handle geometric information for assembly. Boundary graphs represent the faces or surfaces of an assembly part. We show that assembly operations joining two parts together may be modeled simply as the splicing of two boundary graphs representing the two assembly parts. Subgraphs which become detached during the graph splicing process represent surfaces which become inaccessible as a consequence of an assembly operation. These subgraphs are therefore discarded. An added bonus of this approach is that in some instances the feasibility of an assembly operation can be evaluated very quickly by simply scanning the boundary graph for certain geometrical features such as radii of specific surfaces. The modeling of assembly operations and feasibility testing based on boundary graph models advocated here form the basis of a direct or forward assembly planner.

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. 1 Introduction We have previously demonstrated the use of model-based vision for the location, recognition and tracking of vehicles [5][10]. Our techniques have used specially constructed models of 3-D objects, represented as points, lines...

This paper concerns the recovery of intrinsic and extrinsic camera parameters using perspective views of rectangles. Several algorithms are described. The intrinsic parameters (aspect ratio, effective focal length, and the principal point) are determined in closed-form using a minimum of four views. The recovery of the intrinsic parameters and the rotation angles is independent of the size of the rectangles. The absolute translation is recovered by knowing either the length of one side of the rectangle or the area of the rectangle or any other absolute information. The new method for recovering extrinsic parameters is shown to be significantly more robust than the widely used method based on vanishing points. Experiments with both synthetic and real images are presented. 1

A new formulation of a pose refinement technique using ``active'' models is described. An error term derived from the detection of image derivatives close to an initial object hypothesis is linearised and solved by least squares. The method is particularly well suited to problems involving external geometrical constraints (such as the ground-plane constraint). We show that the method is able to recover both the pose of a rigid model, and the structure of a deformable model. We report an initial assessment of the performance and cost of pose and structure recovery using the active model in comparison with our previously reported ``passive'' model-based techniques in the context of traffic surveillance. The new method is more stable, and requires fewer iterations, especially when the number of free parameters increases, but shows somewhat poorer convergence.

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. 1

Linear algorithms are presented for computing the 3-D structure and motion of vehicles in traffic scenes under the ground-plane constraint. Point matches in multiple image frames are used. The algorithms first determine the inter-frame rotation angles, then the point depths, and finally the inter-frame translations. All computations are linear and require no parameter tuning. The algorithms are tested with both synthetic data and routine traffic images. Extensive experimental results are reported that demonstrate the validity of the algorithms.

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. 1 Introduction Any practitioner of 3D vision knows that camera calibration is a time-consuming chore [5]. ...

A new method is described - Struct we from Symmetry (SFS) - in which objects possessing bilateral symmetry may be reconstructed from a single perspective view. The algorithm exploits the groundplane constraint [9, 14] and assumes a camera model, to allow 3D recovery up to scale. The SFS algorithm has been implemented in the form of an interactive tool. The user marks a set of symmetric object points in an image, and a complete facet model description is generated as output. The tool has been extensively used to create models for the vision system developed in the VIEWS project. The algorithm has potential for automatic structure recovery if methods for identifying symmetry points can be improved.

Objects such as vehicles are often constrained to lie on a known plane. The ground-plane constraint reduces the problem of localisation and recognition from 6 dofto 3 dof. A novel algorithm is presented which makes effective use of the ground-plane constraint to derive pose estimates. A form of the generalised Hough transform is used to group evidence from line features, and to identify approximate poses. The single orientation parameter is decoupled from the two location parameters, and treated separately. The method is fast and robust. It copes well with complex outdoor scenes including multiple occluded objects, and image clutter from irrelevant structures. 1

Camera calibration often entails the computation of the perspective transformation matrix. Conventionally, the matrix has been calculated by the standard linear least squares technique. Recently, Faugeras and Toscani have criticised the conventional approach for producing unsatisfactory, even "absurd", solutions, and have proposed an alternative approach. It is shown in this paper that their criticism of the conventional approach is misplaced and misleading. Experimental results demonstrate that Faugeras and Toscani's approach has no advantage over the conventional approach from the practical point of view. In fact, the latter is shown to be superior both in noise robustness and in computational cost. The paper also reports a method to resolve the possible sign ambiguities in the camera parameters computed by existing algorithms. 1

Model based vision allows use of prior knowledge of the shape and appearance of specific objects to be used in the interpretation of a visual scene; it provides a powerful and natural way to enforce the view consistency constraint. A model based vision system has been developed within ESPRIT VIEWS: P2152 which is able to classify and track moving objects (cars and other vehicles) in complex, cluttered traffic scenes. The fundamental basis of the method has been previously reported. This paper presents recent developments which have extended the scope of the system to include (i) multiple cameras, (ii) variable camera geometry, and (iii) articulated objects. All three enhancements have easily been accommodated within the original model-based approach

The authors present a quantitative analysis of the viewpoint consistency constraint (VCC), which is the fundamental principle behind model-based methods for recognizing 3-D objects from 2-D data. It defines a measure of viewpoint consistency error (VCE), based on a formal model of image feature errors. Existing methods for establishing feature correspondences using the VCC are discussed. The poor performance of incremental methods is demonstrated and attributed to the failure to ensure that global consistency improves during search. A more reliable method, viewpoint consistency ascent, which uses the VCE explicitly as a heuristic for a state-space search, is presented. The two algorithms are compared in an experimental study. The approach to quantitative analysis of alternative algorithms is illustrated, which may be applied to model based object recognition more generally.>

A view-independent relational model (VIRM) used in a vision system for recognizing known 3-D objects from single monochromatic images of unknown scenes is described. The system inspects a CAD model from a number of different viewpoints, and a statistical interference is applied to identify relatively view-independent relationships among component parts of the object. These relations are stored as a relational model of the object, which is represented in the form of a hypergraph. Three-dimensional components of the object, which can be associated with extended image features obtained by grouping of primitive 2-D features are represented as nodes of the hypergraph. Covisibility of model features is represented by means of hyperedges of the hypergraph, and the pairwise view-independent relations form procedural constraints associated with the hypergraph edges. During the recognition phase, the covisibility measures allow a best-first search of the graph for acceptable matches.>

Model-based vision techniques, originally developed for the recognition and pose recovery of vehicles in a single image, are used here to track vehicles through a sequence of images. Knowledge of the position of the camera with respect to the ground plane is used to reduce the search space of possible vehicle positions from six dimensions to three. The expected dynamics of vehicles are expressed in a Kalman filter, which predicts the likely poses in successive frames and provides a smoothed description of the vehicles' motion. The notion of equivalence classes defined by a search of the parameter space is developed as an indicator of the performance of the pose-refinement sub-system. The system is illustrated and assessed by using the size of the correct class as a performance measure.>

A hierarchical model of the brain has been developed, based on the ARTTM system, in which nodes represent the structures of interest, and arcs represent 3-D topographical relationships between organs and part-of relationships. The nodes contain attributes of the organs which encode their expected shape and identifying features; they also attach to the recognition rules, which use the measured attributes and relationships to identify the structures. In this paper the image data is pre-processed to create a Laplacian Pyramid by using multiple Difference of Gaussian (DoG) filters (see Figure 2). Most of the reasoning used in the recognition rules is based on zero crossing features of the DoG images, which may be tracked between scales, or between images of adjacent slices.

Uejormable models of elastic structures have been proposed for use in image analysis. The models are based on a minimum energy principle which incorporates both image information and "high level " knowledge of the structures involved. This paper reports a further development of the Finite Element Method (FEM) for use in active contour models. It is shown that parabolic and cubic Finite Elements provide a versatile technique for implementing deformable models. The method is demonstrated on MR and x-ray images of brain sections. 1.

This paper describes the automatic production of three dimensional computer models of arteries, where the originating data is a pair of biplane X-ray angiograms. By developing the process of reconstruction using a hierarchy of generic models, it is possible to build a limited number of plausible reconstructions. There are indications that further research will be able to identify the unique reconslmclion that matches the original structure. The process avoids the extensive human operator intervention and simplifying assumptions used by other researchers. (il.yl

algorithm is reported, and demonstrated in use for interpreting medical images. The SIMD machine acts as a "compute server " to a general-purpose AI environment, which provides and interactive user interface, or controls the server autonomously.

The model has strongly influenced the construction of the design aid both in its functionality and in the specification of the user interface. A designer is able to interact with the tool via a direct manipulation interface which offers a drawing board metaphor. Any part of the evolving hierarchical design representation can be accessed and it supports a range of funtionality from user specified mathematical function to individual electronic components. Multiple versions of all or part of the design are maintained by the system and a history of the design steps and decisions establishes a model of the designers' use of the tool. The initial system has been implemented using some of the facilities provided by the Automatic Reasoning Tool (ART) from Inference Corporation within an object-oriented framework. The limitations of these tools are addressed and the future direction of the work discussed.

Abstract A protocol for file transfer between two computers connected by an asynchronous line and its implementation on a PDP11 are described.