Search

It is more than 10 years since the first tentative experiments in ear biometrics were conducted and it has now reached the “adolescence” of its development towards a mature biometric. Here we present a timely retrospective of the ensuing research since those early days. Whilst its detailed structure may not be as complex as the iris, we show that the ear has unique security advantages over other biometrics. It is most unusual, even unique, in that it supports not only visual and forensic recognition, but also acoustic recognition at the same time. This, together with its deep three-dimensional structure and its robust resistance to change with age will make it very difficult to counterfeit thus ensuring that the ear will occupy a special place in situations requiring a high degree of protection.

Gait recognition has become a popular new biometric in the last decade. Good recognition results have been achieved using different gait techniques on several databases. However, not much attention has been paid to get major questions: how good are biometrics data; how many subjects are needed to cover diversity of population (hypothetical or actual) in gait and how many samples per subject will give good representation of similarities and differences in the gait of the same subject. In this paper we try to answer these questions from the point of view of statistical analysis not only for gait recognition but for other biometrics as well. Though we do not think that we have a whole answer, we content this is the start of the answer.

Gait is a new biometric aimed to recognise a subject by the manner in which they walk. Gait has several advantages over other biometrics, most notably that it is non-invasise and perceivable at a distance when other biometrics are obscured. We present a new area based metric, called gait masks, which provides statistical data intimately related to the gait of the subject. Early results show promising results with a recognition rate of 90% on a small database of human subjects. In addition to this, we show how gait masks can also be used on subjects other than humans to provide information about the gait cycle of the subject.

To recognize people by their gait from a sequence of images, we have proposed a statistical approach which combined eigenspace transformation (EST) with canonical space transformation (CST) for feature transformation of spatial templates. This approach is used to reduce data dimensionality and to optimize the class separability of different gait sequences simultaneously. Good recognition rates have been achieved. Here, we incorporate temporal information from optical flows into three kinds of temporal templates and use them as features for gait recognition in addition to the spatial templates. The recognition performance for four kinds of template features has been evaluated in this paper. Experimental results show that spatial templates, horizontal-flow templates and the combined horizontal-flow and vertical-flow templates are better than vertical-flow templates for gait recognition.

Detecting ellipses is an important computer vision task. In order to overcome the excessive time and storage requirements associated with the Hough transform, some techniques decompose the required parameter space into several sub-spaces. The decomposition is achieved by using geometric constraints which define relative positions between a set of points. This paper shows how positional constraints can be avoided by using the local properties of the position function to decompose the parameter space. This decomposition requires the combination of a pair of edge points and their gradient direction. The decomposition is used to extract ellipses in synthetic and real images.

Detecting geometric primitives in images is one of the basic tasks of computer vision. The Hough transform (HT) and its extensions constitute a popular method for extracting geometric shapes. Primitives on the HT are represented by parametric curves with a number of free parameters. The principal concept of the HT is to define a mapping between an image space and a parameter space. Each edge point in an image is transformed by the mapping to determine cells in the parameter space whose associated parameters are such that the defined primitive passes through the data point. The chosen cells are accumulated and after all the points in an image have been considered, local maxima in the accumulator correspond to the parameters of the specified shape. Because a curve with n parameters requires an n-dimensional parameter space, many applications of the HT concern line and circle detection. In order to overcome the excessive time and space requirements for ellipse extraction, proposed techniques (Yip et al., Pao et al., Yoo and Sethi, Wu and Wang, Ho and Chen) decompose the five dimensional parameter space into several sub spaces of fewer dimensions. The decomposition is achieved by using geometric features which define constraints in the organization of edge data. These constraints include distance and angular relationships which define relative positions between a set of edge points. Hence, the parameters are computed after labelling the points which satisfy the constraints in a computational intensive approach. Here, we show how it is possible to decompose the parameter space, based on the polar definition of an ellipse. Angular information, defined from the variation of a position function, represents local change in the curvature of border points. This information is used to formulate expressions which define an ellipse by including local shape properties. We show that in order to achieve a parameter decomposition (due to the ellipse anisotropy) it is necessary to consider the angular change of the second derivative (tangent angle of the second directional derivative). We compute angular information by taking a pair of points and their gradient direction. This avoids the constraints which define relative position, as required by other approach