Abstract The purpose of this research is to develop a new method for encrypting multiple superimposed or side-by-side images. The process begins by extracting the red, green, and blue channels from each image and converting them into vectors that combine to produce a single image that undergoes an advanced pixel-level Vigenere transform. In the next step, a pseudorandom transition occurs at the nucleotide, followed by a passage to codons for genetic crossover implementation specifically designed for image scrambling. The latter process is controlled by many random tables developed from selected chaotic maps, which ensures a high degree of flexibility and security in the encryption method. To evaluate the effectiveness and security of this innovative multi-image encryption algorithm, extensive simulations were performed using a large number of images randomly selected from the database. The simulation results prove the reliability and robustness of the method.

Abstract

In recent years, brain health has received increasing attention, but conventional acquisition of brain MRI (magnetic resonance imaging) images still suffer from issues such as missing data, artifacts, and high costs, which hinders research and diagnosis. With the application of deep learning in medical image synthesis, low-cost, efficient, and high-quality medical MRI synthesis techniques have become a prominent research focus and have gradually matured. However, traditional methods for synthesizing magnetic resonance imaging (MRI) mostly rely on generative adversarial networks, which require fine-tuning of parameters and learning rates to achieve stringent Nash equilibrium conditions, leading to problems such as gradient explosions and mode collapse. Building upon the latest research in synthetic models DDPM (denoising diffusion probabilistic model), we propose a novel approach for 2D brain MRI image synthesis based on a lightweight denoising diffusion probabilistic model. This method improves the attention module in the denoising diffusion probabilistic model to make it more lightweight. Additionally, we adopt the smooth L1 loss function as a replacement for the traditional mean absolute error (L1 loss) by comparing the error between the 2D brain MRI images with added noise and the real noise for training the model. Finally, we validate the proposed model on the MRI Brain Tumor Classification dataset, demonstrating that it achieves high-quality synthesis results while significantly reducing the parameter count of the DDPM model.

Abstract

Stroke can be mainly categorized into hemorrhagic and ischemic stroke. Intracerebral hemorrhage (ICH) is a subtype of hemorrhagic stroke that is caused due to unconstrained bleeding within the parenchyma of the brain. ICH is one of the major conditions that have a high rate of disease and a high rate of death in a given population. Risk factors for ICH emerged to be age, male gender, hypertension, and intake of alcohol in huge quantities. The frequency of ICH is increased where hypertension is mainly untreated. To improve the prognosis and outcomes of an ICH patient, we need to perform emergent evacuation of blood from the brain parenchyma and prevent edema formation while restricting further neuronal damage due to surgical intervention. Evidence-based guidelines exist for ICH and form the basis for a care framework. The pharmaceutical management of ICH from current literature includes an aggressive reduction in blood pressure, tranexamic acid use, and recombinant activated factor VII administration. In addition, advanced imaging, surgical evacuation of ICH, and minimally invasive surgery techniques for hematoma evacuation could provide great benefits to patients with a large ICH.

Abstract

Radiomics, a quantitative approach to medical imaging, employs computational methods to extract features from the images, revealing hidden characteristics of specific regions. This emerging field leverages advanced techniques to analyze a spectrum of features from modalities, including computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET) scans, aiming to decode tissue characteristics, disease progression, and treatment responses. The radiomics workflow integrates image acquisition, segmentation, feature selection, and data integration, utilizing advanced techniques such as deep learning, machine learning, and data mining. Radiomics demonstrates considerable potential in cancer detection and management, exhibiting high sensitivity and specificity in distinguishing between benign and malignant tumors and predicting outcomes. However, challenges such as imaging protocol variability, overfitting, and standardization requirements impede its broad clinical adoption. Innovations in multi-modal radiomics, deep learning, and genomics integration strive to mitigate these constraints. This review elucidates radiomics’ capabilities, current applications, benefits, challenges, and future directions in oncology.

Abstract

Advancements in Medical Image Segmentation have revolutionized clinical diagnostics and treatment planning. This review explores a wide range of segmentation techniques applied to Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) images, emphasizing their clinical implications and future directions. CT segmentation techniques, including U-Net and its variant nnU-Net, are essential in oncology for precise tumor delineation, in cardiology for coronary artery analysis, and in pulmonology for lung lesion detection. These methods enhance radiotherapy targeting, surgical planning, and overall diagnostic accuracy. The nnU-Net, known for its self-configuring nature, is particularly notable for setting new benchmarks in medical image segmentation tasks. MRI segmentation benefits from superior soft tissue contrast. Techniques like Mask Region-based Convolutional Neural Network (R-CNN) excel in identifying brain lesions, assessing musculoskeletal injuries, and monitoring soft tissue tumors. These methods support detailed visualization of internal structures, improving diagnosis and guiding targeted interventions. U-Net architectures also play a critical role in MRI segmentation, demonstrating high efficacy in various applications such as brain and prostate imaging. A systematic review of the literature reveals performance metrics for various segmentation techniques, such as accuracy, sensitivity, specificity, and processing time. Traditional methods like thresholding and edge detection are contrasted with advanced deep learning and machine learning approaches, highlighting the strengths and limitations of each. The review also addresses methodological approaches, including data collection, analysis, and evaluation metrics. Future prospects include integrating 3D and 4D segmentation, multimodal data fusion, and enhancing AI explain ability. These innovations aim to refine diagnostic processes, personalize treatments, and improve patient outcomes. Clinical applications of these segmentation techniques demonstrate significant advantages in radiology, oncology, and cardiology, though challenges such as data variability and noise persist. Emerging strategies like data augmentation and transfer learning offer potential solutions to these issues. The continuous evolution of medical image segmentation techniques promises to enhance diagnostic accuracy, efficiency, and the personalization of patient care, ultimately leading to better healthcare outcomes.