Papoulis-Gerchberg (PG) algorithm, a technique to extrapolate signals, has attracted many researchers for its lower complexity, higher computational efficiency, and effectiveness. One field that receives these merits is super-resolution, which fuses multiple band-limited scenes to generate a high-resolution image. Most super-resolution methods based on the PG algorithm, however, underperform when input images are seriously degraded by blur, noise, and sampling. The current study addresses the challenges by embedding the PG algorithm into a super-resolution minimization problem. The proposed method is iterative and incorporates a diffusion-driven smoothness prior that updates its regularisation process according to the local image features. This well-crafted prior, which attempts to overcome the super-resolution ill-posedness, provides an automatic interplay between flat and contour regions, and ensures necessary levels of regularisations to generate sharper and detailed images. Results show that the current method outperforms some state-of-the-art super-resolution approaches including those based on total variation. Even more importantly, the authors' method contains a robust noise suppressor that treats comfortably noisy scenes.

This study presents a fast-efficient error concealment method for recovering information related to shape. The proposed technique comprises block classification, edge direction interpolation and filtering interpolation. Missing blocks are classified into four categories: transparent, opaque, edge and isolated blocks. Most of the computation is spent on edge blocks and isolated blocks to maximise the cost and performance tradeoff. For the recovery of edge blocks, the edge slope is computed by referring to the nearest available block, from which the missing shape is interpolated parallel to the edge. Isolated blocks are dealt with using a cascade filter to approximate the actual shape. Experimental results show that the proposed method provides better cost performance in the restoration of shapes than that afforded by comparable algorithms, both in numerical parameters and the resulting shapes. The processing speed is approximately two to three times faster than previous methods and low computational load makes the proposed technique applicable to real-time MPEG-4 systems.

Enhancement of images with weak edges faces great challenges in imaging applications. In this study, the authors propose a novel image enhancement approach based on intuitionistic fuzzy sets. The proposed method first divides an image into sub-object and sub-background areas, and then successively implements new fuzzification, hyperbolisation, and defuzzification operations on each area. In this way, an enhanced image is obtained, where the visual quality of region of interest (ROI) is significantly improved. Several types of images are utilised to validate the proposed method with respect to the enhancement performance. Experimental results demonstrate that the proposed algorithm not only works more stably for different types of images, but also has better enhancement performance, in comparison to conventional methods. This is a great merit of such design for discerning specific ROIs.

Statistical shape models have been widely used to guide the segmentation in an image, thus overcoming noise and occlusions. In this study, the authors present a graph cut-based segmentation framework, in which multiple objects can be segmented. They design a specific multilabel shape prior, which is integrated into the graph cost function. They also want to enforce spatial constraint between the objects. Towards this aim, they propose a local constraint to forbid the inclusion of an object into another, which is enforced in the regularisation term of the graph energy. They apply the authors’ method to cardiac magnetic resonance images, in which left and right ventricles, and the myocardium are segmented and for which encouraging results are obtained.

Active shape models (ASM) showed to have potential for segmenting the right ventricle (RV) in cardiac magnetic resonance images (MRIs). Nevertheless, the large variability and complexity of the RV shape do not allow for concisely capturing all possible shape variations among patients and anatomical cross-sections. Noticeably, the latter increases the number of iterations required to converge to a proper solution and reduces the segmentation accuracy. In this study, the authors propose a new ASM framework that can model the RV shape in short-axis cardiac MRI images. In this framework, the RV contour is split into two simpler segments, septal (SP) and free wall, whose shape variations are independently modelled using two separate (dual) ASM models. The contour splitting is done at the location of the RV insertion points into the SP wall. Further, instead of using the conventional Procrustes method, the RV contours are aligned using the Bookstein coordinate transformation, which uses the RV insertion points as landmarks to linearly align the RV contours. The results from a dataset of 14 patients show that the proposed framework outperforms the conventional ASM framework and can model complex RV shape variation with more accuracy and in less iteration steps.

Detection of coronary artery stenosis, from 3D computed tomography angiography (CTA), is applicable for inspecting heart diseases. In this study, a semi-automatic method is proposed which its stages include 3D CTA pre-processing, vessel enhancement, coronary artery segmentation, centreline extraction, arteries cross-section diameter estimation, and stenosis detection. In contrast to conventional methods, this study is proposed in which the slices are rescaled from original size to smaller size and then returned to the original size, in order to reduce processing time in the centreline extracting step. To optimise final results, a semi-automatic method is proposed to adjust seed points for coronary arteries segmentation using 3D region growing method for reducing human interventions. The authors consider two types of evaluations for stenosis detecting, more than 50%, on 18 real data. The first type is patient-based analysis and the second type is segment-based analysis. In the first type, a sensitivity of 88.89% and a positive predictive value (PPV) of 88.89% are obtained, and in the second type, a sensitivity of 44.2%, and a PPV of 34.27% is achieved. Moreover, the average execution time for stenosis detecting in a 3D CTA is approximately 8.5 min.

Facial landmark detection is a fundamental process included in many face analysis tasks. However, it faces great challenges for the facial images to be detected usually containing large variations, which will deteriorate the detection precision. Among such variations, partial occlusions and pose variations take great effects. In this study, the authors present a mixture of discriminative visibility-aware models (MDVMs) for facial landmark detection, to improve the generalisation ability of the model to occlusion and pose variation. By adopting different structure constrains for different poses as well as selecting different appearance model for the occluded parts, the MDVMs method can efficiently address the problem of partial occlusion and pose variation. Experiment results demonstrate that their proposed MDVMs method outperforms the well-known template-based methods, and can obtain much more accurate and robust facial landmarks detection results under both occlusions and pose variations.

The need for image encryption is constantly on the rise and has led to the emergence of many techniques in literature. In recent years, many chaos-based encryption techniques have developed with more or less success. In this study, we suggest one such approach using two oscillators. Mixed chaotic maps from the Colpitts and Duffing oscillators were used to encrypt images, which helped to increase the key space to 2⁴⁴⁸ ≃ 7.26 × 10¹³⁴. The authors also succeeded to reduce the encryption time by saving the mixed chaotic maps and reusing them as need arose. The proposed system was tested on well-known images like Lena, NebulaM83, Mandrill and Clown.

Human brain is separated into two hemispheres by the mid-sagittal plane (MSP) as bilateral symmetry. Extraction of this symmetry plane from magnetic resonance images is one of the precise processes for diagnosis. The foremost challenge of this work is to analyse the degree of asymmetry between hemispheres. Most of the existing work has analysed primarily on the image intensity to estimate the asymmetry between hemispheres. The present study explores the possibility of the generalised fractal dimensions to measure the asymmetry between hemispheres, in addition multifractal spectra applies to refine the optimal region of interest which characterises the complexity and homogeneity of an object. In order to validate the efficiency of the proposed technique, experimental results are compared with three state-of-the-art methods by the performance evaluation metrics such as yaw angle error and roll angle error. Besides, angular deviation and average deviation of distance between ground truth line and extracted MSP by the developed method is compared.

As digital images are increasing exponentially; it is very attractive to develop more effective machine learning frameworks for automatic image annotation. In order to address the most prominent issues (huge inter-concept visual similarity and huge intra-concept visual diversity) more effectively, an inter-related non-parametric Bayesian classifier training framework to support multi-label image annotation is developed. For this purpose, an image is viewed as a bag, and its instances are the over-segmented regions within it found automatically with an adopted Otsu's method segmentation algorithm. Here firefly algorithm (FA) is utilised to enhance Otsu's method in the direction of finding optimal multilevel thresholds using the maximum variance intra-clusters. FA has high convergence speed and less computation rate as compared with some evolutionary algorithms. By generating blobs, the extracted features for segmented regions, the concepts which are learned by the classifier tend to relate textually to the words which occur most often in the data and visually to the easiest to recognise segments. This allowing the opportunity to assign a word to each object (localised labelling). Extensive experiments on Corel benchmark image datasets will validate the effectiveness of the proposed solution to multi-label image annotation and label ranking problem.

Inspired by the fact that wavelet transform can be written as a classical convolution form, a new linear canonical wavelet transform (LCWT) based on generalised convolution theorem associated with linear canonical transform (LCT) is proposed recently. The LCWT not only inherits the advantages of multi-resolution analysis of wavelet transform (WT), but also has the capability of image representations in the LCT domain. Based on these good properties, the authors propose a novel image watermarking method using LCWT and QR decomposition. Compared with the existing image watermarking methods based on discrete WT or QR, this novel image watermarking method provides more flexibility in the image watermarking. Peak signal-to-noise ratio, normalised cross and structural similarity index measure are used to verify the advantages of the proposed method in simulation experiments. The experiment results show that the proposed method is not only feasible, but also robust to some geometry attacks and image processing attacks.

This study presents a novel method for content-aware image resizing based on the normalised saliency map of an image termed saliency histogram (SH), which provides the probability of a target's presence in the spatial location domain. Motivated by the idea of spreading the histogram of an image on the grey scale known as image histogram equalisation for contrast enhancement, fast SH equalisation (SHE) has been proposed to distribute the content of an image in the resized image domain similarly. Experimental results show that SHE has a tendency to protect salient foreground objects, and it takes only fractions of a second to resize an 1024 × 768 image.