The Human Mental Search (HMS) algorithm is a recent metaheuristic algorithm inspired by the search strategies employed in online auctions. In HMS, each candidate solution is considered a bid, and the algorithm operates in three key stages: mental search, which employs L & eacute;vy flight to explore the neighbourhood of a solution in search of better alternatives; grouping, where candidate solutions are organised into clusters to identify promising areas in search space; and guiding bids towards more promising areas of the solution space. This paper introduces Markov models for the HMS algorithm. The originality of this study lies in providing, for the first time, a theoretical framework that models HMS using Markov chain theory to formally analyse its convergence behaviour. These models provide the theoretical probability distribution for each potential population as the number of generations approaches infinity. In other words, we model the HMS algorithm using the theory of finite element models for continuous optimisers. This approach enables the construction of discrete Markov chain models that simulate the behaviour of HMS in continuous search spaces. The Markov model is then used to investigate and theoretically verify the convergence properties of the HMS algorithm, which are shown to be consistent with previously reported empirical findings.

Arbitrary style transfer (AST), a popular AI-powered photo editing function, aims to strike an optimal balance between content and style injection from two images in order to generate a novel high-fidelity stylised image. Recently, diffusion models have been applied to AST due to their high generation quality as well as flexibility to embed conditions. However, these models are still not satisfactory and may exhibit inferior performance compared to non-diffusion based methods. This is due to the diffusion process not being purposely designed for AST, leading to suboptimal solutions to trade-off content preservation and style embedding. In this paper, we propose ACID-Style, a novel adaptive condition injection diffusion-based AST framework for improved content/style feature injection to address this research challenge. Using two lightweight adapters, a content and a style injection module, and an adaptive injection mechanism, our approach is able to fully exploit a pre-trained stable diffusion model for AST-specific adaptation and our diffusion model thus learns the most effective timing for content and style injection in the diffusion sampling process. Comprehensive evaluations demonstrate that our method achieves superior style transfer performance, both quantitatively and qualitatively, compared to other state-of-the-art style transfer methods.

Weakly supervised semantic segmentation (WSSS) aims to provide accurate pixel-level annotation based on only weak guidance, primarily derived from image-level labels. Recent WSSS methods exploit pseudo-labels generated from improved class activation maps (CAMs) to train a fine-grained classification model for semantic segmentation. However, these pseudo-labels are unreliable because they tend to either miss parts of the objects or include irrelevant regions due to weak guidance from individual images. In this paper, we propose a simple yet effective iterative mix-up strategy, Pseudo-Label-based Mix (PL-Mix), that refines pseudo-labels iteratively, thereby further enhancing WSSS performance. During each iteration, we migrate object regions from pseudo-labels produced in previous steps and render them with new contexts in a mix-up fashion. Due to model consistency enforcement across varied backgrounds and new combinations of multiple objects from enriched image samples, these pseudo-labels progressively become more accurate and reliable. Further enhanced by a masking strategy and a CAM-based earth mover’s distance loss, we achieve state-of-the-art performance on the PASCAL VOC2012 and MS COCO2014 benchmark datasets.

Current high-capacity image steganography methods face challenges in balancing hidden capacity, imperceptibility, and recovery quality. Existing embedding-based image-in-image steganography approaches tend to produce detectable artifacts when hiding multiple images, whereas existing generative methods struggle to conceal full-sized secret images and often generate unrealistic stego images. To address these issues, this paper proposes a novel generative steganography approach that hides multiple secret images in a single realistic generated image. Our main contributions include a meticulously designed autoencoder that compresses and injects secret images into the shallow layer of the generator to increase hidden capacity, a three-stage optimization strategy for stable training to enhance the recovery quality of secret images, and an automatic image selection procedure which explores the advantage of generation diversity to enhance the imperceptibility of stego images. Experimental results demonstrate that our method outperforms embedding-based approaches by achieving higher recovered image quality with a PSNR value of 30.45 dB when concealing four images while maintaining stronger resistance against steganalysis tools, with an accuracy of 50%. Against generative approaches, our method achieves a higher hidden capacity while preserving a superior visual quality of stego images, with a FID of 6.97, surpassing the suboptimal method's FID of 22.72.

Weakly supervised semantic segmentation (WSSS) aims to achieve pixel-level fine-grained image segmentation using only weak guidance such as image-level class labels, thus significantly decreasing annotation costs. Despite the impressive performance showcased by current state-of-the-art WSSS approaches, the lack of precise object localisation limits their segmentation accuracy, especially for pixels close to object boundaries. To address this issue, we propose a novel class activation map (CAM)-based level set method to effectively improve the quality of pseudo-labels by exploring the capability of level sets to enhance the segmentation accuracy at object boundaries. To speed up the level set evolution process, we use Fourier neural operators to simulate the dynamic evolution of our level set method. Extensive experimental results show that our approach significantly outperforms existing WSSS methods on both PASCAL VOC 2012 and MS COCO datasets.

Unauthorised face recognition (FR) systems have posed significant threats to digital identity and privacy protection. To alleviate the risk of compromised identities, recent makeup transfer-based attack methods embed adversarial signals in order to confuse unauthorised FR systems. However, their major weakness is that they set up a fixed image unrelated to both the protected and the makeup reference images as the confusion identity, which in turn has a negative impact on both attack success rate and visual quality of transferred photos. In addition, the generated images cannot be recognised by authorised FR systems once attacks are triggered. To address these challenges, in this paper, we propose a Recoverable Makeup Transferred Generative Adversarial Network (RMT-GAN) which has the distinctive feature of improving its image-transfer quality by selecting a suitable transfer reference photo as the target identity. Moreover, our method offers a solution to recover the protected photos to their original counterparts that can be recognised by authorised systems. Experimental results demonstrate that our method provides significantly improved attack success rates while maintaining higher visual quality compared to state-of-the-art makeup transfer-based adversarial attack methods. Our code and supplementary materials are available on Github.

Determining optimal weights and biases for neural networks is a critical task. While gradient-based methods are widely used for training, they are sensitive to initialisation and susceptible to local optima. Population-based metaheuristics, such as differential evolution (DE), can offer a reliable alternative. Recently, clustering-based DE approaches have been proposed to further improve this process. However, they suffer from increased complexity, particularly with growing network sizes, leading to longer computation times. In this paper, we introduce strategies to reduce the time complexity of clustering-based DE, including clustering in the objective space, a two-tier clustering period, and one-step kappa-means clustering. We select one of the recent training algorithms, C2L-DE, as a representative method to incorporate our proposed strategies, leading to a lightweight version, C2L-DE-Lite. We show that C2L-DE-Lite decreases the complexity from O(root N-pop center dot N-pop center dot d center dot I), where N-pop is the population size, d is the dimensionality, and I is the number of iterations, to O ( N (pop) center dot root N-pop/CP), where CP is the clustering period. This means that the complexity remains constant for increasing sizes of networks. Extensive experiments demonstrate that while significantly reducing time complexity, C2L-DE-Lite maintains similar performance levels.

Determining appropriate weights and biases for feed-forward neural networks is a critical task. Despite the prevalence of gradient-based methods for training, these approaches suffer from sensitivity to initial values and susceptibility to local optima. To address these challenges, we introduce a novel two-level clustering-based differential evolution approach, C2L-DE, to identify the initial seed for a gradient-based algorithm. In the initial phase, clustering is employed to detect some regions in the search space. Population updates are then executed based on the information available within each region. A new central point is proposed in the subsequent phase, leveraging cluster centres for incorporation into the population. Our C2L-DE algorithm is compared against several recent DE-based neural network training algorithms, and is shown to yield favourable performance.

The efficacy of feed-forward multi-layer neural networks relies heavily on their training procedure, where identifying appropriate weights and biases plays a pivotal role. Nonetheless, conventional training algorithms such as backpropagation encounter limitations, including getting trapped in sub-optimal solutions. To rectify these inadequacies, metaheuristic population algorithms are advocated as a dependable alternative. In this paper, we introduce a novel training methodology termed, DDE-OP, which leverages the principles of differential evolution enriched with a division-based scheme and an opposite-direction strategy. Our approach integrates two effective concepts with differential evolution. Initially, the proposed algorithm identifies partitions within the search space through a clustering algorithm and designates the obtained cluster centres to serve as representatives. Subsequently, an updating scheme incorporates these clusters into the current population. Lastly, a quasi-opposite-direction strategy is used to augment search space exploration. Extensive evaluation on diverse classification and approximation tasks demonstrate that DDE-OP surpasses conventional and population-based methodologies.

Event-based semantic segmentation has gained popularity due to its capability to deal with scenarios under high-speed motion and extreme lighting conditions, which cannot be addressed by conventional RGB cameras. Since it is hard to annotate event data, previous approaches rely on event-to-image reconstruction to obtain pseudo labels for training. However, this will inevitably introduce noise, and learning from noisy pseudo labels, especially when generated from a single source, may reinforce the errors. This drawback is also called confirmation bias in pseudo-labeling. In this paper, we propose a novel hybrid pseudo-labeling framework for unsupervised event-based semantic segmentation, HPL-ESS, to alleviate the influence of noisy pseudo labels. Specifically, we first employ a plain unsupervised domain adaptation framework as our baseline, which can generate a set of pseudo labels through self-training. Then, we incorporate offline event-to-image re-construction into the framework, and obtain another set of pseudo labels by predicting segmentation maps on the re-constructed images. A noisy label learning strategy is de-signed to mix the two sets of pseudo labels and enhance the quality. Moreover, we propose a soft prototypical alignment ( SPA) module to further improve the consistency of target domain features. Extensive experiments show that the pro-posed method outperforms existing state-of-the-art methods by a large margin on benchmarks (e.g., +5.88% accuracy, +10.32% mIoU on DSEC-Semantic dataset), and even sur-passes several supervised methods.

Single image novel view synthesis allows the generation of target images with different views from a single input image. Pixel generation methods are one of the main approaches for novel view synthesis, with previous methods typically using the input image to infer the target image in the new view. However, only features from input images in the source view might not be sufficient to generate a good target image, especially when only a single input image is available. In this paper, we fuse features from an input and a warped image to collaboratively generate pixels in the new view, with the warped image as an intermediate output generated by projecting pixels of the input image onto the target view via an estimated depth. Since the estimated depth and the generated warped image are not perfect, errors will be introduced when generating target pixels. To alleviate these and to ensure better channel information between the features from input and warped image, channel attention blocks are employed. In addition, in order to use skip connections for better novel view synthesis results, encoder features in different layers from the input image are transformed to the target view via multi-resolution depths. Here, instead of downsampling a single full-resolution depth to several lower-resolution depths, we adopt a multi-scale depth estimation network to predict multiple depths at different resolutions. Experimental results on benchmark datasets show that our method gives excellent view synthesis results and outperforms other state-of-the-art novel view synthesis methods.

Federated learning (FL) allows multiple participants to collaboratively build deep learning (DL) models without directly sharing data. Consequently, the issue of copyright protection in FL becomes important since unreliable participants may gain access to the jointly trained model. Application of homomorphic encryption (HE) in a secure FL framework prevents the central server from accessing plaintext models. Thus, it is no longer feasible to embed the watermark at the central server using existing watermarking schemes. In this article, we propose a novel client-side FL watermarking scheme to tackle the copyright protection issue in secure FL with HE. To the best of our knowledge, it is the first scheme to embed the watermark to models under a secure FL environment. We design a black-box watermarking scheme based on client-side backdooring to embed a pre-designed trigger set into an FL model by a gradient-enhanced embedding method. Additionally, we propose a trigger set construction mechanism to ensure that the watermark cannot be forged. Experimental results demonstrate that our proposed scheme delivers outstanding protection performance and robustness against various watermark removal attacks and ambiguity attack.

Multi-level image thresholding is a common approach to image segmentation where an image is divided into several regions based on its histogram. Otsu's method is the most popular method for this purpose, and is based on seeking for threshold values that maximise the between-class variance. This requires an exhaustive search to find the optimal set of threshold values, making image thresholding a time-consuming process. This is especially the case with increasing numbers of thresholds since, due to the curse of dimensionality, the search space enlarges exponentially with the number of thresholds. Population-based metaheuristic algorithms are efficient and effective problem-independent methods to tackle hard optimisation problems. Over the years, a variety of such algorithms, often based on bio-inspired paradigms, have been proposed. In this paper, we formulate multi-level image thresholding as an optimisation problem and perform an extensive evaluation of 23 population-based metaheuristics, including both state-of-the-art and recently introduced algorithms, for this purpose. We benchmark the algorithms on a set of commonly used images and based on various measures, including objective function value, peak signal-to-noise ratio, feature similarity index, and structural similarity index. In addition, we carry out a stability analysis as well as a statistical analysis to judge if there are significant differences between algorithms. Our experimental results indicate that recently introduced algorithms do not necessarily achieve acceptable performance in multi-level image thresholding, while some established algorithms are demonstrated to work better.

Scene text image super-resolution (STISR) in the wild has been shown to be beneficial to support improved vision-based text recognition from low-resolution imagery. An intuitive way to enhance STISR performance is to explore the well-structured and repetitive layout characteristics of text and exploit these as prior knowledge to guide model convergence. In this paper, we propose a novel gradient-based graph attention method to embed patch-wise text layout contexts into image feature representations for high-resolution text image reconstruction in an implicit and elegant manner. We introduce a non-local group-wise attention module to extract text features which are then enhanced by a cascaded channel attention module and a novel gradient-based graph attention module in order to obtain more effective representations by exploring correlations of regional and local patch-wise text layout properties. Extensive experiments on the benchmark TextZoom dataset convincingly demonstrate that our method supports excellent text recognition and outperforms the current state-of-the-art in STISR. The source code is available at https://github.com/xyzhu1/TSAN.

Plagiarism detection is a challenging task, aiming to identify similar items in two documents. In this paper, we present a novel approach to automatic plagiarism detection that combines BERT (bidirectional encoder representations from transformers) word embedding, attention mechanism-based long short-term memory (LSTM) networks, and an improved differential evolution (DE) algorithm for weight initialisation. BERT is used to pretrain deep bidirectional representations in all layers, while the pre-trained BERT model can be fine-tuned with only one extra output layer without significant changes in architecture. Deep learning algorithms often use the random weighting method for initialisation, followed by gradient-based optimisation algorithms such as back-propagation for training, making them susceptible to getting trapped in local optima. To address this, population- based metaheuristic algorithms such as DE can be used. We propose an improved DE algorithm with a clustering-based mutation operator, where first a winning cluster of candidate solutions is identified and a new updating strategy is then applied to include new candidate solutions in the current population. The proposed DE algorithm is used in LSTM, attention mechanism, and feed- forward neural networks to yield the initial seeds for subsequent gradient-based optimisation. We compare our proposed model with conventional and population-based approaches on three datasets (SNLI, MSRP and SemEval2014) and demonstrate it to give superior plagiarism detection performance.

Differential evolution (DE) is widely used for global optimisation problems due to its simplicity and efficiency. L-SHADE is a state-of-the-art variant of DE algorithm that incorporates external archive, success-history-based parameter adaptation, and linear population size reduction. L-SHADE uses a current-to-pbest/1/bin strategy for mutation operator, while all individuals have the same probability to be selected. In this paper, we propose a novel L-SHADE algorithm, RWS-L-SHADE, based on a roulette wheel selection strategy so that better individuals have a higher priority and worse individuals are less likely to be selected. Our extensive experiments on the CEC-2017 benchmark functions and dimensionalities of 30, 50 and 100 indicate that RWS-L-SHADE outperforms L-SHADE.

In machine learning, multilayer perceptrons (MLPs) have long been used as one of the most popular and effective classifiers. With training the crucial process, the susceptibility of conventional algorithms such as back-propagation to get stuck in local optima has encouraged many researchers to opt for metaheuristic algorithms instead. In this paper, we propose a novel population-based metaheuristic algorithm for MLP training using Lévy flight distribution (LFD). In our approach, the optimum weights of the network are found via a population of agents moving through the search space either by Lévy flight motions or by random walks. Comparing the results of this algorithm on several datasets from the UCI repository with other population-based metaheuristic algorithms shows excellent results and superiority of the LFD algorithm.

Masi entropy is a popular criterion employed for identifying appropriate threshold values in image thresholding. However, with an increasing number of thresholds, the efficiency of Masi entropy-based multi-level thresholding algorithms becomes problematic. To overcome this, we propose a novel differential evolution (DE) algorithm as an effective population-based metaheuristic for Masi entropy-based multi-level image thresholding. Our ME-GDEAR algorithm benefits from a grouping strategy to enhance the efficacy of the algorithm for which a clustering algorithm is used to partition the current population. Then, an updating strategy is introduced to include the obtained clusters in the current population. We further improve the algorithm using attraction (towards the best individual) and repulsion (from random individuals) strategies. Extensive experiments on a set of benchmark images convincingly show ME-GDEAR to give excellent image thresholding performance, outperforming other metaheuristics in 37 out of 48 cases based on cost function evaluation, 26 of 48 cases based on feature similarity index, and 20 of 32 cases based on Dice similarity. The obtained results demonstrate that population-based metaheuristics can be successfully applied to entropy-based image thresholding and that strengthening both exploitation and exploration strategies, as performed in ME-GDEAR, is crucial for designing such an algorithm.

Foot ulcer is a common complication of diabetes mellitus and, associated with substantial morbidity and mortality, remains a major risk factor for lower leg amputations. Extracting accurate morphological features from foot wounds is crucial for appropriate treatment. Although visual inspection by a medical professional is the common approach for diagnosis, this is subjective and error-prone, and computer-aided approaches thus provide an interesting alternative. Deep learning-based methods, and in particular convolutional neural networks (CNNs), have shown excellent performance for various tasks in medical image analysis including medical image segmentation. In this paper, we propose an ensemble approach based on two encoder-decoder-based CNN models, namely LinkNet and U-Net, to perform foot ulcer segmentation. To deal with a limited number of available training samples, we use pre-trained weights (EfficientNetB1 for the LinkNet model and EfficientNetB2 for the U-Net model) and perform further pre-training using the Medetec dataset while also applying a number of morphological-based and colour-based augmentation techniques. To boost the segmentation performance, we incorporate five-fold cross-validation, test time augmentation and result fusion. Applied on the publicly available chronic wound dataset and the MICCAI 2021 Foot Ulcer Segmentation (FUSeg) Challenge, our method achieves state-of-the-art performance with data-based Dice scores of 92.07% and 88.80%, respectively, and is the top ranked method in the FUSeg challenge leaderboard. The Dockerised guidelines, inference codes and saved trained models are publicly available at https://github.com/masih4/Foot_Ulcer_Segmentation.

Recent years have seen an increasing interest in single image view synthesis. It remains however a challenging task due to the lack of comprehensive colour and depth information from different views. In this paper, we propose a novel view synthesis approach that incorporates a Neural Image Refinement Network (NIRN) and generates both depth and colour images for the target view in an end-to-end manner. The appearance of the colour image greatly benefits from the generated depth image as it provides an intermediate projection relationship for the object in the 3D world. Since the direct application of geometric projection mapping will result in empty regions and/or distortions, our approach proposes to embed a novel refinement network into the view synthesis pipeline for improved performance. Experimental results on three publicly available datasets demonstrate that our NIRN outperforms other state-of-the-art view synthesis methods.