Constructing artificial images of facial expressions

This paper presents a method for construction of artificial images of facial expressions. The proposed fractal-based synthesis procedure called pixel-based correspondence works on 2D images and does not require any depth information. This method can generate artificial images of an object when only a single image is given. Using the proposed method, effective example-based facial analysis systems can be trained and utilised in various applications.

Visual tracking of vehicles using multiresolution analysis and neural network

This paper describes the procedure for detection and tracking of a vehicle from an on-road image sequence taken by a monocular video capturing device in real time. The main objective of such a visual tracking system is to closely follow objects in each frame of a video stream, such that the object position as well as other geometric information are always known. In the tracking system described, the video capturing device is also moving. It is a challenge to detect and track a moving vehicle under a constantly changing environment coupled to real time video processing. The system suggested is robust to implement under different illuminating conditions by using the monocular video capturing device. The vehicle tracking algorithm is one of the most important modules in an autonomous vehicle system, not only it should be very accurate but also must have the safety of other vehicles, pedestrians, and the moving vehicle itself. In order to achieve this an algorithm of multi resolution technique based on Haar basis functions were used for the wavelet transform, where a combination of classification was carried out with the multilayer feed forward neural network. The classification is done in a reduced dimensional space, where principle component analysis (PCA) dimensional reduction technique has been applied to make the classification process much more efficient. The results show the effectiveness of the proposed methodology.

Detection of vehicles with monolithic classifier vis-à-vis a boosted cascaded classifier

This paper describes the comparison of accuracy and performance of two machine learning approaches for visual object detection and tracking vehicles, from an on-road image sequence. The first is a neural network based approach. Where an algorithm of multi resolution technique based on Haar basis functions was used to obtain an image with different scales. Thereafter a classification was carried out with the multilayer feed forward neural network. Principle Component Analysis (PCA) technique was used as a dimension reduction technique to make the classification process much more efficient. The second approach is based on boosting which also yields very good detection rates. In general, boosting is one of the most important developments in classification methodology. It works by sequentially applying a classification algorithm to reweighed versions of the training data, followed by taking a weighted majority vote of the sequence of classifiers thus produced. For this work, a strong classifier was trained by the adaboost algorithm. The results of comparing the two methodologies visà-vis shows the effectiveness of the methods that have been used.

Joint learning and dictionary construction for pattern recognition

We propose a joint representation and classification framework that achieves the dual goal of finding the most discriminative sparse overcomplete encoding and optimal classifier parameters. Formulating an optimization problem that combines the objective function of the classification with the representation error of both labeled and unlabeled data, constrained by sparsity, we propose an algorithm that alternates between solving for subsets of parameters, whilst preserving the sparsity. The method is then evaluated over two important classification problems in computer vision: object categorization of natural images using the Caltech 101 database and face recognition using the Extended Yale B face database. The results show that the proposed method is competitive against other recently proposed sparse overcomplete counterparts and considerably outperforms many recently proposed face recognition techniques when the number training samples is small.

Boosting performance for 2D linear discriminant analysis via regression

Two Dimensional Linear Discriminant Analysis (2DLDA) has received much interest in recent years. However, 2DLDA could make pairwise distances between any two classes become significantly unbalanced, which may affect its performance. Moreover 2DLDA could also suffer from the small sample size problem. Based on these observations, we propose two novel algorithms called Regularized 2DLDA and Ridge Regression for 2DLDA (RR-2DLDA). Regularized 2DLDA is an extension of 2DLDA with the introduction of a regularization parameter to deal with the small sample size problem. RR-2DLDA integrates ridge regression into Regularized 2DLDA to balance the distances among different classes after the transformation. These proposed algorithms overcome the limitations of 2DLDA and boost recognition accuracy. The experimental results on the Yale, PIE and FERET databases showed that RR-2DLDA is superior not only to 2DLDA but also other state-of-the-art algorithms.

Efficient tensor based face recognition

This paper addresses the limitation of current multilinear PCA based techniques, in terms of pro- hibitive computational cost of testing and poor gen- eralisation in some scenarios, when applied to large training databases. We define person-specific eigen-modes to obtain a set of projection bases, wherein a particular basis captures variation across light- ings and viewpoints for a particular person. A new recognition approach is developed utilizing these bases. The proposed approach performs on a par with the existing multilinear approaches, whilst sig- nificantly reducing the complexity order of the testing algorithm.

Face recognition using kernel ridge regression

In this paper, we present novel ridge regression (RR) and kernel ridge regression (KRR) techniques for multivariate labels and apply the methods to the problem of face recognition. Motivated by the fact that the regular simplex vertices are separate points with highest degree of symmetry, we choose such vertices as the targets for the distinct individuals in recognition and apply RR or KRR to map the training face images into a face subspace where the training images from each individual will locate near their individual targets. We identify the new face image by mapping it into this face subspace and comparing its distance to all individual targets. An efficient cross-validation algorithm is also provided for selecting the regularization and kernel parameters. Experiments were conducted on two face databases and the results demonstrate that the proposed algorithm significantly outperforms the three popular linear face recognition techniques (Eigenfaces, Fisherfaces and Laplacianfaces) and also performs comparably with the recently developed Orthogonal Laplacianfaces with the advantage of computational speed. Experimental results also demonstrate that KRR outperforms RR as expected since KRR can utilize the nonlinear structure of the face images. Although we concentrate on face recognition in this paper, the proposed method is general and may be applied for general multi-category classification problems.

The use of injury surveillance databases to identify emerging injury hazards

Among the many valuable uses of injury surveillance is the potential to alert health authorities and societies in general to emerging injury trends, facilitating earlier development of prevention measures. Other than road safety, to date, few attempts to forecast injury data have been made, although forecasts have been made of other public health issues. This may in part be due to the complex pattern of variance displayed by injury data. The profile of many injury types displays seasonality and diurnal variance, as well as stochastic variance. The authors undertook development of a simple model to forecast injury into the near term. In recognition of the large numbers of possible predictions, the variable nature of injury profiles and the diversity of dependent variables, it became apparent that manual forecasting was impractical. Therefore, it was decided to evaluate a commercially available forecasting software package for prediction accuracy against actual data for a set of predictions. Injury data for a 4-year period (1996 to 1999) were extracted from the Victorian Emergency Minimum Dataset and were used to develop forecasts for the year 2000, for which data was also held. The forecasts for 2000 were compared to the actual data for 2000 by independent t-tests, and the standard errors of the predictions were modelled by stepwise hierarchical multiple regression using the independent variables of the standard deviation, seasonality, mean monthly frequency and slope of the base data (R = 0.93, R² = 0.86, F_{(3, 27)} = 55.2, p < 0.0001). Significant contributions to the model included the SD (β = 1.60, p < 0.001), mean monthly frequency (β =  - 0.72, p < 0.002), and the seasonality of the data (β = 0.16, p < 0.02). It was concluded that injury data could be reliably forecast and that commercial software was adequate for the task. Variance in the data was found to be the most important determinant of prediction accuracy. Importantly, automated forecasting may provide a vehicle for identifying emerging trends.

Road-sign identification using ensemble learning

Ensemble learning that combines the decisions of multiple weak classifiers to from an output, has recently emerged as an effective identification method. This paper presents a road-sign identification system based upon the ensemble learning approach. The system identifies the regions of interest that are extracted from the scene into the road-sign groups that they belong to. A large road-sign image dataset is formed and used to train and test the system. Fifteen groups of road signs are chosen for identification. Five experiments are performed and the results are presented and discussed.

Detection and classification of road signs in natural environments

An automatic road sign recognition system first locates road signs within images captured by an imaging sensor on-board of a vehicle, and then identifies the detected road signs. This paper presents an automatic neural-network-based road sign recognition system. First, a study of the existing road sign recognition research is presented. In this study, the issues associated with automatic road sign recognition are described, the existing methods developed to tackle the road sign recognition problem are reviewed, and a comparison of the features of these methods is given. Second, the developed road sign recognition system is described. The system is capable of analysing live colour road scene images, detecting multiple road signs within each image, and classifying the type of road signs detected. The system consists of two modules: detection and classification. The detection module segments the input image in the hue-saturation-intensity colour space, and then detects road signs using a Multi-layer Perceptron neural-network. The classification module determines the type of detected road signs using a series of one to one architectural Multi-layer Perceptron neural networks. Two sets of classifiers are trained using the Resillient-Backpropagation and Scaled-Conjugate-Gradient algorithms. The two modules of the system are evaluated individually first. Then the system is tested as a whole. The experimental results demonstrate that the system is capable of achieving an average recognition hit-rate of 95.96% using the scaled-conjugate-gradient trained classifiers.

Knowledge on road information in sub-urban lane detection via multiple cue integration

Detection of lane boundaries of a road based on the images or video taken by a video capturing device in a suburban environment is a challenging task. In this paper, a novel lane detection algorithm is proposed without considering camera parameters; which robustly detects lane boundaries in real-time especially for sub-urban roads. Initially, the proposed method fits the CIE L*a*b* transformed road chromaticity values (that is a* and b* values) to a bi-variate Gaussian model followed by the classification of road area based on Mahalanobis distance. Secondly, the classified road area acts as an arbitrary shaped region of interest (AROI) in order to extract blobs resulting from the filtered image by a two dimensional Gabor filter. This is considered as the first cue of images. Thirdly, another cue of images was employed in order to obtain an entropy image. Moreover, results from the color based image cue and entropy image cue were integrated following an outlier removing process. Finally, the correct road lane points are fitted with Bezier splines which act as control points that can form arbitrary shapes. The algorithm was implemented and experiments were carried out on sub-urban roads. The results show the effectiveness of the algorithm in producing more accurate lane boundaries on curvatures and other objects on the road.

Multi-scale local pattern co-occurrence matrix for textural image classification

Textural image classification technologies have been extensively explored and widely applied in many areas. It is advantageous to combine both the occurrence and spatial distribution of local patterns to describe a texture. However, most existing state-of-the-art approaches for textural image classification only employ the occurrence histogram of local patterns to describe textures, without considering their co-occurrence information. And they are usually very time-consuming because of the vector quantization involved. Moreover, those feature extraction paradigms are implemented at a single scale. In this paper we propose a novel multi-scale local pattern co-occurrence matrix (MS_LPCM) descriptor to characterize textural images through four major steps. Firstly, Gaussian filtering pyramid preprocessing is employed to obtain multi-scale images; secondly, a local binary pattern (LBP) operator is applied on each textural image to create a LBP image; thirdly, the gray-level co-occurrence matrix (GLCM) is utilized to extract local pattern co-occurrence matrix (LPCM) from LBP images as the features; finally, all LPCM features from the same textural image at different scales are concatenated as the final feature vectors for classification. The experimental results on three benchmark databases in this study have shown a higher classification accuracy and lower computing cost as compared with other state-of-the-art algorithms.

A fuzzy clustering approach to content-based image retrieval

In this study, an interactive Content-Based Image Retrieval (CBIR) system that allows searching and retrieving images from databases is designed and developed. Based on the fuzzy c-means clustering algorithm, the CBIR system fuses color and texture features in image segmentation. A technique to form compound queries based on the combined features of different images is devised. This technique allows users to have a better control on the search criteria, thus a higher retrieval performance can be achieved. A database consisting of skin cancer imagery is used to demonstrate the applicability of the CBIR system.

Infrared face recognition: a comprehensive review of methodologies and databases

Automatic face recognition is an area with immense practical potential which includes a wide range of commercial and law enforcement applications. Hence it is unsurprising that it continues to be one of the most active research areas of computer vision. Even after over three decades of intense research, the state-of-the-art in face recognition continues to improve, benefitting from advances in a range of different research fields such as image processing, pattern recognition, computer graphics, and physiology. Systems based on visible spectrum images, the most researched face recognition modality, have reached a significant level of maturity with some practical success. However, they continue to face challenges in the presence of illumination, pose and expression changes, as well as facial disguises, all of which can significantly decrease recognition accuracy. Amongst various approaches which have been proposed in an attempt to overcome these limitations, the use of infrared (IR) imaging has emerged as a particularly promising research direction. This paper presents a comprehensive and timely review of the literature on this subject. Our key contributions are (i) a summary of the inherent properties of infrared imaging which makes this modality promising in the context of face recognition; (ii) a systematic review of the most influential approaches, with a focus on emerging common trends as well as key differences between alternative methodologies; (iii) a description of the main databases of infrared facial images available to the researcher; and lastly (iv) a discussion of the most promising avenues for future research. © 2014 Elsevier Ltd.