Richard T. Shann

Efficient method for finding the position of object boundaries to sub-pixel precision

Abstract A new class of algorithms is described for the analysis boundaries in a discrete image. The simplest algorithm of this class accepts as input an initial estimate for the boundary position and, after a process of iterative refinement, outputs a more accurate estimate. The method is an extension of known methods for edge detection which are based on Gaussian filtering. Instead of using a discrete filter and exhaustive evaluation, the filter output is computed only at isolated points. These points are selected by a numerical optimization routine to converge on the feature of interest. Because in general the points do not coincide with pixels the filter must be reconstructed between pixels from the discrete image data. This method permits the measurements to be made to sub-pixel accuracy without the need for a mathematical model of the boundary. The reliability of the method is discussed in terms of the pixel size (sampling error) and the size and proximity of clutter relative to the size of the filter. The accuracy is related to the filter size. Possible applications include non-contact measurement, and an example is given.

Detection and characterization of carboniferous foraminifera for content-based retrieval from an image database

Carboniferous Foraminifers are a specific type of microfossil which are manifest in plane sections of rock and are used by geologists for dating rock samples. The images contain a high degree of visual noise and currently must be interpreted by human experts. We are studying the classification problem in the context of intelligent image databases. Here we present a technique for automatic identification of microfossil structures and for classification of the structures according to which type of 3-D section they represent. This is achieved by using: (1) A specialized filter to detect local curves in the gray level image data; and (2) Hough transform processing of the resulting feature point vectors. An interesting aspect of our approach is that the processing of the features is not embedded in a program but is instead specified using a visual query language. This allows us to experiment quickly with different types of grouping criteria. The detection performance of our system is comparable with that of a trained geologist. We store the information obtained in a database together with the raw image data. The system can then present the user with only those images which contain structures of interest.

Novel approach to boundary finding

Abstract A new class of algorithms is described for the analysis of boundaries in a discrete image. The methods are based on well known methods of Gaussian filtering coupled with the use of numerical methods. Instead of the conventional discrete filter and exhaustive evaluation by convolution, the filter output is computed only at points of interest. In addition, derivative information is gathered at these points which permits rapid convergence onto features of interest. Speed of computation and sub-pixel precision are features of these methods. The directional derivative of Gaussian and Laplacian of Gaussian are discussed, and possible applications indicated.

An Efficient and Robust Local Boundary Operator

We present the Curved Iterative Boundary Locator (CIBL), which is a new algorithm for determining the position and local radius of curvature of a boundary which is based upon the IBL algorithm [1]. Both the IBL and CIBL use the grey level image directly, rather than an edge image, and this distinguishes them from conventional robust model fitting techniques. The performance of the CIBL is evaluated for a realistic image domain and the results are compared with data obtained from robust Least Squares ellipse fitting. We conclude that the CIBL and its variants provide a powerful technique for robust analysis, particularly in the area of industrial inspection where dimension measurement to high precision is often required.

Manchester visual query language

We report a database language for visual retrieval which allows queries on image feature information which has been computed and stored along with images. The language is novel in that it provides facilities for dealing with feature data which has actually been obtained from image analysis. Each line in the Manchester Visual Query Language (MVQL) takes a set of objects as input and produces another, usually smaller, set as output. The MVQL constructs are mainly based on proven operators from the field of digital image analysis. An example is the Hough-group operator which takes as input a specification for the objects to be grouped, a specification for the relevant Hough space, and a definition of the voting rule. The output is a ranked list of high scoring bins. The query could be directed towards one particular image or an entire image database, in the latter case the bins in the output list would in general be associated with different images. We have implemented MVQL in two layers. The command interpreter is a Lisp program which maps each MVQL line to a sequence of commands which are used to control a specialized database engine. The latter is a hybrid graph/relational system which provides low-level support for inheritance and schema evolution. In the paper we outline the language and provide examples of useful queries. We also describe our solution to the engineering problems associated with the implementation of MVQL.

A curvature sensitive filter and its application in microfossil image characterisation

A new class of oriented, curvature sensitive filters are introduced. These filters provide a low-level detection facility for noisy curves without a prior edge extraction stage. The application of these filters to the detection of Carboniferous Foraminifers (a type of microfossil found in plane rock sections) is described. A symbolic representation of the detected curves is stored in a database which is then queried to recover the required structures. We show that the curves identified by the filter correspond to salient features of the microfossil evidence in the image.

Local Method for Curved Edges and Corners

A new method of detecting features in Gaussian smoothed images is described. Applied to the simple case of marking the Canny edge pixels the method gives an improved response at comers where traditional methods have problems. Moreover, the method permits marking of more sophisticated features at essentially no extra cost. Additional information available includes local curvature estimates for each edge, direct (local) identification of occluded edges and marking and characterising comers on edges. The performance of the method on real images is compared with the Plessey-Harris corner detector. We find that the quality and sensitivity of these comer detection methods are similar, with the new method giving information on the orientation and opening angle of the corners. Finally it is shown how ellipse detection in noisy images is made possible using this curvature information.

A Novel Approach to Boundary Finding.

A new class of algorithms is described for the analysis of boundaries in a discrete image. The methods are based on well known methods of Gaussian filtering coupled with the use of numerical methods. Instead of the conventional discrete filter and exhaustive evaluation by convolution the filter output is computed only at points of interest. In addition derivative information is gathered at these points which permits rapid convergence onto features of interest. Speed of computation and sub-pixel precision are features of these methods. The directional derivative of Gaussian andLaplacian of Gaussian are discussed and possible applications indicated.