D. E. Rees

Multi-Object Tracking in Video

This paper reports on tracking of multiple objects using color histogram backprojection and motion cues. Four tasks which facilitate this are discussed. The first is an adaptive color histogram backprojection (which builds upon the works of Swain and Ballard) and its application to tracking of multiple objects in video sequences. The second task is designing efficient fast blob detectors for selecting regions of interest in video sequences. The third is motion detection based on color histogram backprojection. Achieving these tasks led to multi-objects tracking. Various video sequences were used to demonstrate effective tracking of multiple objects. Notably, we created an interactive multiple objects tracker (CLICK-IT) which in its present form is set at three objects but can be extended easily. CLICK-IT (CSIRO Laboratory for Imaging by Content and Knowledge?Interactive Television) is a PC-based system which provides the user with an intelligent highlighter pen for sports action replay. It is intended as a truly interactive improvement on the drawing pad technology currently used for video annotation in sports broadcasting. The system uses computer vision techniques to focus attention and track particular objects (player(s), ball, horse(s), ?) and semi-automatically annotate the dynamic scene. This paper describes the system including the user interface, the tracking technology based on color and motion information, and system performance evaluation in applications to surveillance-like sequences, running, rugby league football, basketball and soccer. Finally, video scene detection based on color histogram is discussed.

An improved synergetic algorithm for image classification

Abstract A major application of pattern recognition technology is in industrial manufacturing. In this paper, we develop a synergetic algorithm for pattern recognition which is based purely on the appearance of the object, without reference to a CAD model of the object, making the technique generic and flexible. In particular, we apply this algorithm to the problem of classifying an object into a number of user-defined aspects, which is an important problem in robotic manipulation of objects. The technique is fast and can be trained using a non-iterative, deterministic training scheme which will find a zero-error solution on a training set, if such a solution exists.

The CLICK (CSIRO Laboratory for Imaging by Content and Knowledge) Security Demonstrator

CSIRO Telecommunications and Industrial Physics is developing digital image processing tools using object oriented technology on a PC platform, with an emphasis on the study of image content and the application of domain knowledge. This paper summarises progress made in the following areas: (a) image database browsing, (b) real-time face recognition and verification, and (c) still image and video compression. The focus of the paper is on security applications of these technologies which have been integrated into the CLICK (CSIRO Laboratory for Imaging by Content and Knowledge) Security Demonstrator.

Pattern recognition techniques and the measurement of solar magnetic fields

Measuring vector magnetic fields in the solar atmosphere using the profiles of the Stokes parameters of polarized spectral lines split by the Zeeman effect is known as Stokes Inversion. This inverse problem is usually solved by least-squares fitting of the Stokes profiles. However least-squares inversion is too slow for the new generation of solar instruments (THEMIS, SOLIS, Solar-B, ...) which will produce an ever-growing flood of spectral data. The solar community urgently requires a new approach capable of handling this information explosion, preferably in real-time. We have successfully applied pattern recognition and machine learning techniques to tackle this problem. For example, we have developed PCA-inversion, a database search technique based on Principal Component Analysis of the Stokes profiles. Search is fast because it is carried out in low dimensional PCA feature space, rather than the high dimensional space of the spectral signals. Such a data compression approach has been widely used for search and retrieval in many areas of data mining. PCA-inversion is the basis of a new inversion code called FATIMA (Fast Analysis Technique for the Inversion of Magnetic Atmospheres). Tests on data from HAOs Advanced Stokes Polarimeter show that FATIMA isover two orders of magnitude faster than least squares inversion. Initial tests on an alternative code (DIANNE - Direct Inversion based on Artificial Neural NEtworks) show great promise of achieving real-time performance. In this paper we present the latest achievements of FATIMA and DIANNE, two powerful examples of how pattern recognition techniques can revolutionize data analysis in astronomy.

Learning in a self-organising pattern formation system

Abstract In this paper we implement learning of a set of unlabelled images using the analogy between pattern formation and pattern learning. Our algorithm is clearer, more robust, and of lower dimension than comparable synergetic algorithms in the literature.

Explicit inversion: an approach to image analysis

Abstract Image analysis can be expressed as an inverse problem. Given an image, which is the output of some complicated and possibly unknown function, our goal is to estimate the parameters of that function. Formally, at least, the solution to the problem can be found by inverting the function which produced the image. In practice, this inversion requires two major elements; a feature extractor and a parameter estimator. While there has been much research into these two elements, they are generally designed separately from one another. In this paper we introduce an approach to image analysis founded on the belief that these two elements should be designed as a pair. We label our approach ‘explicit inversion’, because it allows us to replace the problem of implicitly inverting an unknown, possibly high-dimensional function, with that of explicitly inverting a known, low-dimensional function. As a result we achieve major time reductions over the standard approaches while achieving comparable accuracy.