This study presents a new dual-tree complex wavelet transform (DT-CWT)-based illumination normalisation approach for face recognition under varying lighting conditions. The method consists of three steps. First, the DT-CWT-based edge detection method is proposed which can obtain estimation for facial feature edges in different directionality and resolution level. Second, the DT-CWT-based denoising model is employed to obtain the multi-scale illumination invariant structures in the logarithm domain. Finally, by combining the obtained illumination invariant features and edge estimation information, the enhanced facial features are obtained which have more discriminating power for variable lighting face recognition. The effectiveness of the method is validated in comparative performance against many classical illumination compensation methods using the YaleB database and the CMU PIE database.

In real-world scenes, vehicles are frequently overlapped by other objects and various backgrounds. In this study, an effective method to detect such vehicles, especially those partially occluded by nearby vehicles or other objects is presented. The authors have developed a statistical approach to generate occlusion hypothesis and a new hypothesis verification method. To verify occlusion hypothesis, the verification method utilises geometric and likelihood information. In this way, both vehicle–background and vehicle–vehicle occlusions can be detected. No additional occlusion-specific training is required. In addition, a median filter is applied to eliminate the noise in the patch scoring, and a union-find algorithm is used to find the connected positive region in the binary map. A synthesised occlusion dataset is created to test the performance, and the experimental results on popular benchmarks indicate that the proposed method is effective and robust in recognising partially occluded vehicles.

The authors introduce the concept of a spline resampling in the particle filter to deal with high accuracy and sample impoverishment. The resampling is usually based on a linear transformation on the weights of the particles, so it affects the filtering accuracy. The spline resampling consists of two parts: the spline transformation of weights and the spread transformation of states. The former is based on a spline transformation on the weights of the particles to obtain highly accurate particle filtering, and the latter is based on a point spread transformation on states of particles to prevent sample impoverishment due to a decline in the diversity of hypothesis after resampling. Two transformations are sequentially implemented to incorporate with each other. Then, the authors propose a global transition model in the particle filter, which takes account of the background variation caused by the camera motion, to decrease error from real object position. The authors test the performance of their spline resampling and the global transition model in the particle filter in an object-tracking scenario. Experimental results demonstrate that the particle filter with the spline resampling and the global transition model has promising discriminative capability in comparison with others.

Inverse texture synthesis (ITS) plays an important role in many computer vision applications. It aims at generating small compaction images to capture significant features of textures, from which arbitrarily large images with similar textures can be re-synthesised. This study presents a fast approach to ITS with two algorithms developed in hierarchical wavelet packet (WP) trees: the modified inverse texture synthesis (MITS) algorithm, which extends the ITS algorithm into the WP domain, and the wavelet-packet-tree-based cropping (WPTC) algorithm for the initialisation of MITS. Experimental results show that the combination of WPTC and MITS termed the image compaction in wavelet packet trees algorithm outperforms the ITS algorithm in terms of the ‘peak signal-to-noise ratio’ and computation time.

Based on a graph-theoretic concept of a cluster, dominant sets clustering has been shown to be an attractive clustering algorithm with many useful properties. In this study, the authors conduct a comprehensive study of related issues in dominant sets clustering, in an endeavour to explore the potential of this algorithm and obtain the best clustering results. Specifically, they empirically investigate how similarity parameters, similarity measures and game dynamics influence the dominant sets clustering results. From experiments on eight datasets, they conclude that distance-based similarity measures perform evidently better than cosine and histogram intersection similarity measures potentially, and they need to find the best-performing similarity parameter to make use of this advantage. They then study the effect of similarity parameter on dominant sets clustering results and induce the range of the best-performing similarity parameters. Furthermore, they find that the recently proposed infection and immunisation dynamics performs better than the replicator dynamics in most cases while being much more efficient than the latter. These observations are helpful in applying dominant sets clustering to practical problems, and also indicate directions for further improvement of this algorithm.

Sparse representations for classification (SRC) are considered a relevant advance to the biometrics field, but are particularly sensitive to data misalignments. In previous studies, such misalignments were compensated for by finding appropriate geometric transforms between the elements in the dictionary and the query image, which is costly in terms of computational burden. This study describes an algorithm that compensates for data misalignments in SRC in an implicit way, that is, without finding/applying any geometric transform at every recognition attempt. The authors' study is based on three concepts: (i) sparse representations; (ii) projections on orthogonal subspaces; and (iii) discriminant locality preserving with maximum margin projections. When compared with the classical SRC algorithm, apart from providing slightly better performance, the proposed method is much more robust against global/local data misalignments. In addition, it attains performance close to the state-of-the-art algorithms at a much lower computational cost, offering a potential solution for real-time scenarios and large-scale applications.

Recently, much research has been carried out on dimensionality reduction techniques that summarise a large set of features into a smaller set, leading to much less redundancy. Among the various non-linear dimensionality reduction techniques, Isomap and local linear embedding are the most popular techniques. However, these techniques require optimisation on the size of the neighbourhood set, reducing their computational efficiency. To overcome this limitation, a computational efficient algorithm, which is not dependent on size of the neighbourhood set, has been proposed in this study. In the proposed algorithm, morphological mapping (Morphmap) has been used for visualising an image in three dimensions. Then neighbourhood graph construction is done using intensity attenuation function. The pixels in the layer which comes under the angle of divergence can form a part of group vectors and other vectors are removed. In the proposed method, neighbourhood graph construction is not dependent on size of neighbourhood set and removes all the redundant data that reduces complexity to a great extent. Various facial expression datasets have been used to evaluate the performance of the proposed method. Finally, improvement in recognition accuracy with reduced number of computations using the Morphomap algorithm has been shown in the study.

Osteoporosis is a bone disease characterised by a reduction in bone mass, resulting in an increased risk of fractures. Doctors need the bone mineral density (BMD) measurements of vertebral bodies in order to diagnose and treat osteoporosis. The authors' objective is to segment the VBs as accurately as possible and hence to increase the accuracy of the BMD measurements and fracture analysis. Three pieces of information (intensity, spatial interaction and shape) are modelled to optimise a probabilistic energy functional. A universal shape prior, which is modelled using the cervical, thoracic and lumbar spinal regions, is proposed. Volumetric computed tomography data sets with various challenges are used in this study. The authors classify data sets based on some features related to the anatomy, imaging modality and level of the bone health. The proposed framework is one of only a few reported in the literature tested on the data obtained from different imaging devices. The experimental results reveal that the proposed method is robust under various noise levels, less variant to the initialisation and faster than existing vertebrae segmentation reports in the literature.

Sparse representation (or sparse coding) has been applied to deal with frontal face recognition. Two representative methods are the sparse representation-based classification (SRC) and the collaborative representation-based classification (CRC), in which the query face image is represented by a sparse linear combination of all the training samples. The difference between SRC and CRC is that the L ₁-norm constraint of coding is employed in the former to guarantee the sparse property, while the L ₂-norm constraint is utilised in the latter. In this paper, we propose a novel loose L _1/2 regularised sparse representation (SR) for face recognition, named L _1/2 classification (LHC), which is inspired by L _1/2 regularisation. Additionally, an iterative Tikhonov regularisation (ITR) is proposed to solve LHC efficiently compared with the original algorithm. Using ITR, the balance between the collaborative representation (CR) and the SR can be tuned by the iterations. Attributed to the sparser L _1/2 regularisation and the iterative solution mechanism, a better performance can be achieved by LHC. Extensive experiments on three benchmark face databases demonstrated that LHC is more effective than the state-of-the-art SR-based methods in dealing with frontal face recognition.

Tensor voting is a computational framework that addresses the problem of perceptual organisation. It was designed to convey human perception principles into a unified framework that can be adapted to extract visually salient elements from possibly noisy or corrupted images. The original formulation featured some concerns that made it difficult or impractical to be applied directly. Therefore, several partial or total theoretical reformulations or augmentations have been proposed. These almost parallel publication were presented in different directions, with different priorities and even in a different notation. Thus, the authors observed the need for a coherent description and comparison of the different proposals. This work, after describing the original approach of tensor voting, reviews and explains a number of high impact theoretical modifications in a self-contained manner and including possible future directions of work. The authors have selected and organised a number of formulations and unified the way the problem is addressed across the different proposals. The aim of this study is to contribute with a modern comprehensive source of information on the theoretical aspects of tensor voting.

This study presents a local image descriptor based on spectral embedding. Specifically, the spectra of line graph are used to represent image edges, corners and edge points with big curvature. The authors theoretically analyse and experimentally verify that the spectra of line graph are robust to noise and are invariant to rotation and linear intensity changes. Based on such a fact, some local image descriptors are constructed using the spectra of line graph. Comparative experiments demonstrate the effectiveness of the proposed descriptor and its superiority to some state-of-the-art descriptors under image rotation, image blur, viewpoint change, illumination change, JPEG compression and noise.

Saliency detection has attracted considerable attention, and numerous approaches aimed at locating meaningful regions in images have been presented. Nevertheless, accurate saliency detection algorithms remain in urgent demand. Many algorithms work well when dealing with simple images, but work poorly with complex images that contain small-scale and high-contrast structures. Moreover, most existing local and global regional saliency detection methods measure image saliency through region contrast. Such measurement is achieved by directly computing the difference between non-adjacent regions. In this study, the authors introduce a new perspective for evaluating region contrast. We propose a novel multi-scale saliency region detection method by optimising the shortest path of two non-adjacent regions in the colour space and by measuring the region contrast from different scales. The final saliency maps indicate that the proposed method can work well with images containing small patches, but with high contrast. The proposed approach can also make the foreground significantly more uniform. Experimental results on three public benchmark datasets show that the proposed method achieves better precision–recall curve than some state-of-the-art methods.

Symbol understanding is a pre-requisite for reading and interpretation of any map. Thus, any topographic map-based automated system needs to understand generic symbols associated with the set of maps. The objective of this study was to develop an automated system for the understanding of Survey of India topographic map symbols. The system has been developed making use of shape analysis adopting a complex-valued chain coding method for representation of the (exterior) boundary of the symbol. Fourier discrete transform and autocorrelation function have been used for shape descriptions. Classification and recognition have been implemented through a template matching method based on similarity measures. The system has been trained with 150 samples of 20 Indian topographic symbols and tested for 200 samples of each of the 20 symbols extracted from sample maps. Experimental results showed that the proposed method has an overall recognition rate of 84.68% as well as the improved mean average precision of symbol recognition. However, understanding of interconnected and/or crowded symbols may be taken up as future scope of this work.