This study presents an advanced modification of the You Only Look Once version 7 (YOLOv7) model named Gallbladder YOLOv7 (GB-YOLOv7). GB-YOLOv7 integrates a Normalization-based Attention Module (NAM) and a Global Attention Mechanism (GAM) into the backbone and head architecture. Several image preprocessing methods are also employed, including median filtering and Contrast-Limited Adaptive Histogram Equalization (CLAHE). The framework includes three attention mechanism-based models, including Coordinate and Global Attention Mechanism (CordGAM-YOLOv7), Dual Global Attention Mechanism (DualGAM-YOLOv7), and Normalization-based Attention Module YOLOv7 (NAM-YOLOv7), enabling a meticulous comparative analysis with GB-YOLOv7. Results demonstrate its superior performance across all metrics compared to both traditional and newer YOLO versions: achieving a Recall of 91.3% (vs YOLOv8's 78.1% and YOLOv9's 84.8%), a Mean Average Precision of 94.0%, and a Specificity of 96.2% (vs YOLOv11's 90.6%). GB-YOLOv7 also shows significant improvements in Matthews Correlation Coefficient (MCC) (72.7% vs YOLOv7's 67.3%) and F1-score (90.0% vs You Only Look Once version 9 ( YOLOv9)'s 81.4%), while maintaining greater parameter efficiency (24.34M vs YOLOv7's 36.58M), showcasing its potential as a cutting-edge tool for more effective gallbladder cancer detection.

In healthcare, it is essential for any Large Language Model (LLM)-generated output to be reliable and accurate, particularly in cases involving decision-making and patient safety. However, the outputs are often unreliable in such critical areas due to the risk of hallucinated outputs from the LLMs. To address this issue, we propose a fact-checking module that operates independently of any LLM, along with a domain-specific summarization model designed to minimize hallucination rates. Our model is fine-tuned using Low-Rank Adaptation (LoRA) on the MIMIC-III dataset and is paired with the fact-checking module, which uses numerical tests for correctness and logical checks at a granular level through discrete logic in natural language processing (NLP) to validate facts against electronic health records (EHRs). We trained the LLM on the full MIMIC-III dataset. For evaluation of the fact-checking module, we sampled 104 summaries, extracted them into 3786 propositions, and used these as facts. The fact-checking module achieves a precision of 0.8904, a recall of 0.8234, and an F1-score of 0.8556. Additionally, the LLM summary achieves a ROUGE-1 score of 0.5797 and a BERTScore of 0.9120 for summary quality.

Parkinson’s disease (PD) is one of the fastest‐growing neurodegenerative disorders, where timely diagnosis is essential for optimizing treatment. In this study, we created a radiomics–MDS‐UPDRS, a robust dataset by integrating DaTscan SPECT radiomics data with the clinical characteristics of MDS‐UPDRS collected from Parkinson’s progression markers initiative (PPMI) to monitor dopamine depletion in the striatum (caudate and putamen) and allow classification and progression analysis of PD. To construct the dataset, the striatum was segmented using a modified K‐means clustering algorithm, extracting 25 radiomics features combined with 59 clinical features. In addition, linear discriminant analysis was used to select 22 significant characteristics, and a four‐way feature selection method was used to identify 30 significant clinical features, resulting in a refined set of 52. Classification with machine learning models improved performance after LDA, achieving over 91% accuracy. We evaluated feature behavior across six PD severity stages and four clinical visits for progression analysis. The clinical features of MDS‐UPDRS were more sensitive to changes in the severity of the initial PD. At the same time, the integrated dataset, radiomics–MDS‐UPDRS, provided more balanced insights, showing a progression of 33.30%–83.30% and 36.36%–45.50% from the first visit to the fourth visit among the clinical and radiomics features and a progression of 73.33%–96.67% and 13.64%–54.55% between the minimal vs mild and minimal vs very severe stage. Our analysis also revealed practical links between progression features and real‐life scenarios, which highlights the practical value of our study for clinical decision‐making.

Personalized federated learning (PFL) is a promising technique for tackling data heterogeneity in federated learning systems. Recently, Bayesian neural networks (BNNs) have been introduced into the PFL framework to enable uncertainty quantification and improve performance in data-scarce settings. Despite these advantages, existing BNN-based PFL methods face two key challenges in practical applications. First, in real-world scenarios, client heterogeneity often arises in the form of group-wise variation, which cannot be adequately captured by a single shared distribution as assumed in prior work. Second, existing methods rely on deterministic or stochastic approximation techniques for posterior inference, which lead to substantial computational and memory overhead, hindering their scalability and deployment. To address these limitations, we propose DBFed, a novel BNN-based PFL framework from a distribution-based perspective. DBFed introduces group-specific distributions to better model the structural heterogeneity commonly observed in federated settings. Moreover, DBFed employs a rank-1 parameterization technique to map uncertainty from the weight space to a low-dimensional subspace, significantly reducing the computational and memory overhead. Theoretically, we establish the effectiveness of the rank-1 parameterization approach. Empirically, extensive experiments on diverse datasets demonstrate that DBFed consistently outperforms alternative PFL baselines in a heterogeneous setting.

Accurate and early identification of gastrointestinal (GI) lesions is crucial for treating and preventing GI diseases, including cancer. Automated computer-aided diagnosis methods can assist physicians in early and accurate detection. Video classification of GI endoscopic videos is challenging due to the complexity and variability of visual data. This research proposes a novel method for classifying GI diseases using endoscopic videos. Leveraging the public HyperKvasir dataset, we applied preprocessing algorithms to enhance GI frames by removing noise and artifacts with morphological opening and closing techniques, ensuring high-quality visuals. We addressed dataset imbalance by proposing a novel algorithm. Our hybrid model, Atrous Spatial Pyramid Pooling with Swin Transformer (ASPPST), combines advanced Convolutional Neural Networks and the Swin Transformer to classify GI videos into 30 distinct classes. We incorporated Gradient-Class Activation Mapping (Grad-CAM) in ASPPST's final layer to improve model explainability. The proposed model achieved 97.49 % accuracy in classifying 30 GI diseases, outperforming other transfer learning models and transformers by 8.04 % and 3.99 %, respectively. It also demonstrated a precision of 97.80 %, recall of 97.77 %, and an F1 score of 97.75 %, showcasing robustness across metrics. The high accuracy of ASPPST makes it suitable for real-world use, delivering fewer errors and more precise results in GI endoscopy video classification. Our approach advances artificial intelligence (AI) in computer vision and deep learning for biomedical engineering applications. Grad-CAM integration enhances transparency, boosting clinician trust and adoption of AI tools in diagnostic workflows. • Efficient preprocessing methods were employed to remove noise and artifacts, enabling the extraction of significant features. • To address data imbalance, an under-sampling strategy was proposed. • A highly accurate hybrid deep learning model was developed to classify gastrointestinal diseases into 30 distinct classes.

ABSTRACT Mobile cloud computing (MCC) combines the portability of mobile devices with cloud data centers to provide advanced services. MCC serves us in various ways in our daily lives, including multimedia streaming, mobile gaming, mobile corporate apps, and data‐intensive mobile applications such as augmented reality and virtual reality. Among the several challenges involved in achieving the best performance for this service, job scheduling emerges as a particularly critical one. User satisfaction, cloud service provider requirements, user priority, cloud provider's resource limitation, user deadline, cloud provider's energy consumption, etc., are the main constraints while maintaining job scheduling in mobile cloud computing. To improve the quality of service (QoS) and achieve the effectiveness of job scheduling, we have proposed a multi‐objective model to balance the situation between user gratification and the cloud service provider's demand. To optimize the cost efficiency of the virtual machine, two types of jobs represent unconstrained and constrained jobs in the cloud data center. The shortest execution first scheduling (SEFS) algorithm is applied for the unconstrained job, and efficient deadline and priority job scheduling (EDPS) algorithm is applied for the constrained job. Our proposed algorithm improves the performance of the existing state‐of‐the‐art algorithms. Reducing the execution time of jobs and minimizing the resource consumption of cloud providers are the improvements of our proposed algorithm.

Unmanned aerial vehicles (UAVs) have been employed for a variety of inspection and monitoring tasks, including agricultural applications and search and rescue (SAR) in remote areas. However, traditional monitoring methods tend to focus on optimizing one aspect. This study aims to propose a complete framework by integrating advanced methods to provide a robust and accurate path coverage solution. The combination of edge detection and area decomposition with a pathfinding algorithm can improve the overall performance. An effective edge detection model is developed that simultaneously detects the boundary and segments the area of interest (AOI) from the aerial land images and provides precise area mapping of the area. An intuitive grid decomposition with grid‐to‐graph mapping improves the flexibility of the area decomposition and ensures maximal coverage and safe operation routes for the UAVs. Finally, a robust modified simulated annealing (MSA) algorithm is introduced to determine the shortest path coverage route. The performance of the proposed methodology is tested on aerial imagery. Area decomposition ensures that there are no gaps in the AOI during the coverage planning. The MSA algorithm obtains the minimum length cost, charge consumption cost, and minimum number of turns to cover the area. It is shown that the integration of these techniques enhances the performance of the coverage path planning (CPP). A comparison of the proposed approach with benchmark algorithms further demonstrates its effectiveness. This study contributes to creating a complete CPP application for UAVs, which may assist with precision agriculture as well as safe and secure rescue operations.

ABSTRACT Colorectal cancer (CRC) is characterized by the uncontrollable growth of cancerous cells within the rectal mucosa. In contrast, colon polyps, precancerous growths, can develop into colon cancer, causing symptoms like rectal bleeding, abdominal pain, diarrhea, weight loss, and constipation. It is the leading cause of death worldwide, and this potentially fatal cancer severely afflicts the elderly. Furthermore, early diagnosis is crucial for effective treatment, as it is often more time‐consuming and laborious for experts. This study improved the accuracy of CRC multi‐class classification compared to previous research utilizing diverse datasets, such as NCT‐CRC‐HE‐100 K (100,000 images) and CRC‐VAL‐HE‐7 K (7,180 images). Initially, we utilized various image processing techniques on the NCT‐CRC‐HE‐100 K dataset to improve image quality and noise‐freeness, followed by multiple feature extraction and selection methods to identify prominent features from a large data hub and experimenting with different approaches to select the best classifiers for these critical features. The third ensemble model (XGB‐LightGBM‐RF) achieved an optimum accuracy of 99.63% with 40 prominent features using univariate feature selection methods. Moreover, the third ensemble model also achieved 99.73% accuracy from the CRC‐VAL‐HE‐7 K dataset. After combining two datasets, the third ensemble model achieved 99.27% accuracy. In addition, we trained and tested our model with two different datasets. We used 80% data from NCT‐CRC‐HE‐100 K and 20% data from CRC‐VAL‐HE‐7 K, respectively, for training and testing purposes, while the third ensemble model obtained 98.43% accuracy in multi‐class classification. The results show that this new framework, which was created using the third ensemble model, can help experts figure out what kinds of CRC diseases people are dealing with at the very beginning of an investigation.

Almost 85% of companies polled said they were looking into anomaly detection (AD) technologies for their industrial image anomalies. The present problem concerns detecting anomalies often occupied by redundant data. It can be either in images or in videos. Finding a correct pattern is a challenging task. AD is crucial for various applications, including network security, fraud detection, predictive maintenance, fault diagnosis, and industrial and healthcare monitoring. Many researchers have proposed numerous methods and worked in the area of AD. Multiple anomalies and considerable intraclass variation make industrial datasets tough. Further, research is needed to create robust, efficient techniques that generalize datasets and detect anomalies in complex industrial images. The outcome of this study focuses on various AD methods from 2019 to 2023. These techniques are categorized further into machine learning (ML), deep learning (DL), and federated learning (FL). It explores AD approaches, datasets, technologies, complexities, and obstacles, emphasizing the requirement for effective detection across domains. It explores the results achieved in various ML, DL, and FL AD methods, which helps researchers explore these techniques further. Future research directions include improving model performance, leveraging multiple validation techniques, optimizing resource utilization, generating high‐quality datasets, and focusing on real‐world applications. The paper addresses the changing environment of AD methods and emphasizes the importance of continuing research and innovation. Each ML and DL AD model has strengths and shortcomings, concentrating on accuracy and performance while applying quality parameters for evaluation. FL provides a collaborative way to improve AD using distributed data sources and data privacy.

Abstract Cervical cancer is a prevalent disease affecting the cervix cells in women and is one of the leading causes of mortality for women globally. The Pap smear test determines the risk of cervical cancer by detecting abnormal cervix cells. Early detection and diagnosis of this cancer can effectively increase the patient’s survival rate. The advent of artificial intelligence facilitates the development of automated computer-assisted cervical cancer diagnostic systems, which are widely used to enhance cancer screening. This study emphasizes the segmentation and classification of various cervical cancer cell types. An intuitive but effective segmentation technique is used to segment the nucleus and cytoplasm from histopathological cell images. Additionally, handcrafted features include different properties of the cells generated from the distinct cervical cytoplasm and nucleus area. Two feature rankings techniques are conducted to evaluate this study’s significant feature set. Feature analysis identifies the critical pathological properties of cervical cells and then divides them into 30, 40, and 50 sets of diagnostic features. Furthermore, a graph dataset is constructed using the strongest correlated features, prioritizes the relationship between the features, and a robust graph convolution network (GCN) is introduced to efficiently predict the cervical cell types. The proposed model obtains a sublime accuracy of 99.11% for the 40-feature set of the SipakMed dataset. This study outperforms the existing study, performing both segmentation and classification simultaneously, conducting an in-depth feature analysis, attaining maximum accuracy efficiently, and ensuring the interpretability of the proposed model. To validate the model’s outcome, we tested it on the Herlev dataset and highlighted its robustness by attaining an accuracy of 98.18%. The results of this proposed methodology demonstrate the dependability of this study effectively, detecting cervical cancer in its early stages and upholding the significance of the lives of women.

Abstract The automatic assessment of perceived image quality is crucial in the field of image processing. To achieve this idea, we propose an image quality assessment (IQA) method for blurriness. The features of gradient and singular value were extracted in this method instead of the single feature in the traditional IQA algorithms. According to the insufficient size of existing public image quality assessment datasets to support deep learning, machine learning was introduced to fuse the features of multiple domains, and a new no-reference (NR) IQA method for blurriness denoted Feature fusion IQA(Ffu-IQA) was proposed. The Ffu-IQA uses a probabilistic model to estimate the probability of each edge detection blur in the image, and then uses machine learning to aggregate the probability information to obtain the edge quality score. After that uses the singular value obtained by singular value decomposition of the image matrix to calculate the singular value score. Finally, machine learning pooling is used to obtain the true quality score. Ffu-IQA achieves PLCC scores of 0.9570 and 0.9616 on CSIQ and TID2013, respectively, and SROCC scores of 0.9380 and 0.9531, which are better than most traditional image quality assessment methods for blurriness.

Abstract The increasing incidence of skin cancer necessitates advancements in early detection methods, where deep learning can be beneficial. This study introduces SkinNet-14, a novel deep learning model designed to classify skin cancer types using low-resolution dermoscopy images. Unlike existing models that require high-resolution images and extensive training times, SkinNet-14 leverages a modified compact convolutional transformer (CCT) architecture to effectively process 32 × 32 pixel images, significantly reducing the computational load and training duration. The framework employs several image preprocessing and augmentation strategies to enhance input image quality and balance the dataset to address class imbalances in medical datasets. The model was tested on three distinct datasets—HAM10000, ISIC and PAD—demonstrating high performance with accuracies of 97.85%, 96.00% and 98.14%, respectively, while significantly reducing the training time to 2–8 s per epoch. Compared to traditional transfer learning models, SkinNet-14 not only improves accuracy but also ensures stability even with smaller training sets. This research addresses a critical gap in automated skin cancer detection, specifically in contexts with limited resources, and highlights the capabilities of transformer-based models that are efficient in medical image analysis.

Deep Learning Neural Networks offer a powerful tool to process visual data and to make decisions, but a limitation is its black box nature which offers low transparency to human examiners. This hurdle presents a particular challenge in using Neural Networks in safety-critical systems, which require high performance and transparency such as medical diagnosis of life-threatening diseases. This paper seeks to build and test a neural network for grading gliomas of comparable function and complexity to others in the field, then to apply Data Visualisation techniques to render the internal workings of the NN more understandable to a human observer. The purpose is to develop a system that can classify brain tumors into low-grade gliomas (LGG) and high-grade gliomas (HGG), to aid with diagnosis and prognosis The Brain Tumor Segmentation Challenge 2020 (BraTS2020) data set was used, with data categorised based on a combination of grade assigned in BraTS2020, and the labels in the segmentation data. As some categories are over-represented, methods were employed to ensure a better balance between different categories. Data augmentation was used to expand the limited number of scans in the BraTS2020. A 3D convolution neural network (CNN) was constructed to grade gliomas. With the method developed in this paper, an accuracy of 94.1% was achieved. A newly devised method to visually represent the weights of a convolution is explored. These graphs, called ‘weight graphs’ allow convolutions to be condensed into a visual medium. The weight graph is designed for easy visual interpretation of the weights assigned within a particular convolution. To overcome the limitations of weight graphs, an alternate graph was devised, called a balance graph, because it shows the overall balance of weights in a kernel, allowing for a quick impression of what effect a single kernel has. It is demonstrated that Balance Graphs improve the accessibility and transparency of the of the weights in convolution layers.

Nuclei segmentation and classification of hematoxylin and eosin-stained histology images is a challenging task due to a variety of issues, such as color inconsistency that results from the non-uniform manual staining operations, clustering of nuclei, and blurry and overlapping nuclei boundaries. Existing approaches involve segmenting nuclei by drawing their polygon representations or by measuring the distances between nuclei centroids. In contrast, we leverage the fact that morphological features (appearance, shape, and texture) of nuclei in a tissue vary greatly depending upon the tissue type. We exploit this information by extracting tissue specific (TS) features from raw histopathology images using the proposed tissue specific feature distillation (TSFD) backbone. The bi-directional feature pyramid network (BiFPN) within TSFD-Net generates a robust hierarchical feature pyramid utilizing TS features where the interlinked decoders jointly optimize and fuse these features to generate final predictions. We also propose a novel combinational loss function for joint optimization and faster convergence of our proposed network. Extensive ablation studies are performed to validate the effectiveness of each component of TSFD-Net. The proposed network outperforms state-of-the-art networks such as StarDist, Micro-Net, Mask-RCNN, Hover-Net, and CPP-Net on the PanNuke dataset, which contains 19 different tissue types and 5 clinically important tumor classes, achieving 50.4% and 63.77% mean and binary panoptic quality, respectively. The code is available at: https://github.com/Mr-TalhaIlyas/TSFD.

Breast Cancer is one of the leading causes of death worldwide. Early detection is very important in increasing survival rates. Intensive research is therefore done to improve early detection of such cancers through the use of available technology. This includes various image processing techniques andgeneral machine learning. However, the reported accuracy for many of these studies was often not at the desirable level. Deep Learning based techniques are a promising approach for the early detection of Breast Cancer. We have therefore done a comparative analysis of seven Deep Learning techniques applied to the Wisconsin Breast Cancer (Diagnostic) Dataset. Long Short Term Memory (LSTM) and Gated Recurrent Unit (GRU) were proven to be the most effective algorithms as these have demonstrated good results for the majority of performance indicators used in this study, including an accuracy of over 99 percent.

A significant proportion of people around the world are currently suffering from the harmful effects of diabetes and a considerable number of them not being identified at an early stage. Over time this may result in serious health problem such as blindness and kidney failure. To accurately classify the disease, different machine learning (ML) approaches can be utilized. In this context, four separate ML algorithms, namely Gradient Boosting (GB), Support Vector Machine (SVM) AdaBoost (AB), and Random Forest (RF) are evaluated using the Pima Indians diabetes dataset, first with based on all features, then to the features selected with the Minimal Redundancy Maximal Relevance (MRMR) Feature Selection (FS) approach. Seven different types of performance evaluation metrics were computed with a 10-fold cross-validation (CV) approach. Computational complexity is also evaluated. The best results were obtained with the Random Forest approach, achieving an accuracy of 99.35%.

The research involves the study of auditory-evoked potentials (AEPs) recorded using electroencephalography (EEG) from human subjects. The study aims to mathematically model how the brain responds to the audio stimulus, which involves the identification of differences in the AEPs by simulating binaural hearing. This is achieved by transmitting auditory stimulus in-phase in both ears, and tones which are 180 degrees out of phase in each ear. The study creates a range of models with the aim to determine the type and order of the models which provide the best fit to the AEPs, and to analyze the differences between the homo-phasic and anti-phasic models. The work discovered that multi-input single-output (MISO) transfer function models are able to fit the AEPs. Tenth-order models provide optimal mathematical fit; these models produced significantly greater fit than lower order models while higher order models produce minimal improvement. The addition of zeros also produced insignificant improvement upon the mathematical fit. About 75-95% of mathematical fits were achieved across all subjects. Analysis of the pole-zero plots suggest that the pole pairs with frequencies greater than 125 rad/s are more damped for the trials using homo-phasic auditory stimulus compared with models generated for trials using anti-phasic stimulus. This suggests that if the brain is processing binaural hearing, then the high-frequency poles in 10-pole MISO transfer functions should have low damping.