Image Processing And Its Applications, 1999. Seventh International Conference on (Conf. Publ. No. 465)

The colours in images are confounded by the colour of the illuminant: a yellow light leads to colours that are more yellow than they ought to be. Colour constancy algorithms attempt to estimate and correct for the colour of the illuminant. Unfortunately existing colour constancy algorithms do not work very well. The reason for their failure is easy to understand: different collections of reflectan... View full abstract»

Low level image processing is the first step in many computer vision applications. Edge detection is one of the most useful low level processes and it has been widely studied for grey level images. In some cases, a grey-level edge detector is not rich enough to describe changes that are present in the images and the use of colour information and colour edges may be a valuable tool for some problem... View full abstract»

Internet multimedia conferencing applications have traditionally used a separate tool approach. In contrast, this paper presents a framework for media-engines to facilitate inter-media integration, without introducing undue performance compromises. This work represents a general trend away from individual media tools that are difficult to integrate, and monolithic application software that is diff... View full abstract»

A new adaptive learning method is developed for the enhancement of subsampled colour images. The system use edge information in the luminance channel to aid the reconstruction of the chrominance signals such that edges between the luminance and the chrominance channels are well aligned. A technique exploiting the luminance information effectively is developed. The method is computationally efficie... View full abstract»

Adequate image quality models should be applicable over a wide range of different types of pictures and distortions. Filtering and masking in conformity with known properties of the human visual system are important preprocessing components of such models. A first approach of an image quality model of this kind was given by Zetzsche and Hauske (1989). Further investigations by Xu (1997) showed tha... View full abstract»

In this note we describe a new class of zero-tree image compression algorithms which are under evaluation for international image compression standards. A wavelet or multiwavelet image decomposition can be viewed as a tree structured set of coefficients giving a hierarchical data structure for representing the transformed image. Some of these tree structures may have all the coefficients zero, suc... View full abstract»

This paper describes the high-speed vision system and the novel image processing algorithm developed for seam tracking applications. This system was used in the gantry robot constructed in Liverpool University for high-speed edge trimming of fabric embroideries. Parallelism between image grabbing, image processing and communication with PC for enhancing the performance of the vision system is also... View full abstract»

This work presents a method of extracting texture features from a Gabor transform data block and the application of these features for texture segmentation by clustering feature vectors. For a given image, 16 Gabor features using Gabor kernels with four scales and four orientations are computed. Filtered images are computed by using a Gabor filter bank on a 32×32 windowed neighborhood for ea... View full abstract»

A new methodology is proposed for supervised fusion-classification of multispectral images based on fuzzy set theory. The method is suited to mapping land-cover in a highly complex landscape. As the fuzzy set theory is intrinsically suited for dealing with the mixed pixels problem and is able to represent ill-defined classes in a natural way, the proposed method overcomes the drawbacks of conventi... View full abstract»

One of the key issues in the field of 3D object recognition is the provision of an automatic and complete modeling system for 3D objects. Since a single range finder can only provide the 2½D range image from a single viewpoint, the complete 3D description of the object is not attainable directly. However, several 2½D range maps can be extracted from different viewpoints and integrate... View full abstract»

In this paper we present an algorithm for fish otolith growth ring detection using a multiagent system. Up to now, the identification of growth rings, for age estimation, is routinely achieved by human readers, but this task is tedious and depends on the reader's subjectivity. One of the major problems encountered during an automatic contour detection is the lack of ring continuity perception. Act... View full abstract»

Neural networks have been previously used to solve invariant pattern classification problems. Several solutions can be used. The first one consists of a pure neural solution. The neural network learns by itself the pre-processing required and extracts local information. This method gives good results but is very difficult to tune because it needs many parameters which are gray level of image pixel... View full abstract»

This paper presents a new method for the coding of image sequences based on examining pixel intensities in the temporal domain. It discusses how to compress and decompress a series of images using the proposed method and what type of image sequences are most applicable to this type of coding. The paper concludes by comparing results from the linear prediction technique with standard MPEG-1 coding,... View full abstract»

Recently developed video coding schemes such as MPEG-4 permit arbitrarily-shaped objects to be encoded as separate entities. We describe a shape coding algorithm that adapts the MPEG-4 arbitrary shape coding strategies to a variable block size framework. The new technique, when used in conjunction with motion compensation schemes employing quadtrees as a coding mechanism, successfully integrates t... View full abstract»

This paper presents a high level programming environment for FPGA-based image processing. The system is based on an FPGA coprocessor whose high level instruction set is based on the operators of image algebra. The paper describes the user's programming interface, and outlines the approach to generating FPGA architectures dynamically for the image coprocessor View full abstract»

Gabor wavelet filter responses used as graph nodes in labelled graph matching have previously been shown to give high accuracy for head pose estimation from a single two-dimensional image. However, the robustness of this approach to additive or convolutive noise, and illumination level has not been assessed systematically. This paper presents an experimental evaluation of the combination of Gabor ... View full abstract»

This paper is concerned with the process of creating an image of the topology of the seabed. At present, many sonar systems exist which create images of representations of the seabed. They are representations since the signals are essentially formed in time and do not represent exactly the seabed. Furthermore, processing algorithms for object detection and texture segmentation assume that the seab... View full abstract»

This paper reports on the development and validation of an image analysis system for the study of an important organism in the life science laboratory. The system uses an active shape model to describe the organism and the form it is allowed to take during normal locomotion. Preliminary results are given in tracking and measurement View full abstract»

The distribution of colours in an image has proven to be very useful for object recognition. Building on Swain's colour indexing (1991), colour distributions are now an integral part of many recognition schemes. This is not to say that colour alone suffices but rather that colour is one important cue that aids recognition. In this paper we look at colour distribution based recognition from a rathe... View full abstract»

This paper presents and evaluates clustering techniques for the training of radial basis function neural networks (Moody and Dark 1989; Broomhead and Lowe 1988). The clustering techniques define the centers of the radial basis functions used by these networks. Therefore, the main purpose is to verify the influence of different clustering techniques in the performance of RBF networks. The K-means (... View full abstract»

This paper describes what is believed to be the first model of the distortions produced by median filters that is able to predict essentially exact numerical values of relevant boundary shifts. High accuracy is attained by taking full account of the actual discrete processes that are involved in the median computation, rather than by assuming a continuum of pixels within the operating window View full abstract»

We have designed an efficient parallel algorithm for performing 3D image reconstruction. In this framework, we have considered 3-D image to be reconstructed from a series of 2-D images, produced using ultrasonography, computer tomography, etc. The paper discusses a general parallel algorithm for 3D image reconstruction over CRCW, CREW and EREW PRAM models. We have developed efficient implementatio... View full abstract»

Some methods for segmenting MRI data, especially those which employ multi-spectral data, can be very time-consuming. Hence, it is desirable to reduce the dimensionality of the multi-spectral data by eliminating regions which are of no interest. When, for example, the objective is the segmentation of the brain, it is clear that a considerable amount of the image could be discarded without compromis... View full abstract»

This paper presents the problems encountered when mapping prototype algorithms onto a hardware platform. The application area for the developed hardware is the enhancement of objects within complex scenes to aid object tracking. The aim of this paper is to evaluate the expected performance of the mapped algorithms using a state of the art commercial-off-the-shelf (COTS) pipeline processing board. ... View full abstract»

Stereoscopic vision is the set of techniques that try to recover three-dimensional information from two or more views of a scene. The usual technique consists of computing correspondences between points in both images in order to be able to calculate their depth, thus obtaining their three-dimensional location. Usually, the correspondences are represented by a disparity map from which we can obtai... View full abstract»