The growing demand for high-quality video processing in real-time applications demands efficient denoising techniques that can operate swiftly while maintaining visual fidelity. Conventional approaches often struggle to balance these competing requirements, especially when dealing with high-resolution video streams or resource-constrained environments. This study aims to develop methods for accelerating video data denoising using deep convolutional neural networks while maintaining acceptable output quality. We selected the popular FastDVDNet denoising network, which operates on a sliding window principle, as our baseline for comparison and a starting point for our research. This paper proposes several modifications of FastDVDNet that significantly enhance computational efficiency. We introduce four key optimizations: caching intermediate denoising results, reducing intermediate channels in input block, simplifying convolutional blocks, and halving the number of channels. We evaluated these modifications on the Set8 dataset and compared the results with the original model at various noise levels. Finally, we introduce LiteDVDNet, a fine-tuned version of FastDVDNet model that achieves the optimal balance between processing speed, and denoising performance. We developed two model variants: LiteDVDNet-32, which is 3× faster than the original model with only 0.18 dB average PSNR reduction, and the more lightweight LiteDVDNet-16, which delivers a 5× speed improvement at the cost of 0.61 dB average PSNR reduction. 

Image contrast is very important visual characteristics that will considerably improve the appearance of the image. In this paper image contrast is to be enhanced optimally to accurately portray all the data in the image using nature inspired meta-heuristic algorithms. Algorithms have been devised and proposed to enhance the contrast of low contrast images in this work. Poor image contrast caused by a low-quality capturing device, biased user experience, and an unsuitable environment setting during image capture is the main problem encountered during the image enhancement process. Histogram Equalization (HE), a frequently used technique for contrast enhancement, typically produces images with unwanted artifacts, an unnatural appearance, and washed-out appearances. The degree of enhancement is beyond the control of the global HE. The quality of an image is crucial for human comprehension, making image contrast enhancement (ICE) a crucial pre-processing stage in image processing and analysis. In the current study, the Pelican Optimization Algorithm, a contemporary meta-heuristic (MH) algorithm influenced by nature, is used as the foundation for the grayscale image contrast enhancement (GICE) approach (POA). The comparison of proposed method with existing contrast enhancement techniques has been done on the basis of standard image quality metrics. The proposed algorithm performance on standard test image and Kodak dataset demonstrates that total image contrast and information provided in the image are both greatly improved by the suggested POA-based image enhancement technique.

Effective communication is paramount in election campaigns, and slogans are crucial in conveying messages and eliciting voter sentiment. This paper introduces CRFVReC (Conditional Random Field with Variable-Length Receptive Fields), a statistical modeling technique to categorize and generate election campaign slogans by analyzing their sentiment. It is a novel approach for sentiment-based slogan generation and analysis in election campaigns. The reason was to choose datasets that precisely capture voter sentiment from a variety of sources such as social media (SM) posts, public comments, and news articles. The datasets were meticulously chosen to encompass a broad spectrum of sentiments and issues that are pertinent to voters. The CRFVReC model was set up to maximize the performance of sentiment classification and slogan generation. Modifying parameters such as the length of the receptive field to match the length of slogans enhanced the model's adaptability and increased its accuracy. Utilizing Conditional Random Fields (CRFs), CRFVReC classifies election campaign slogans into optimistic and pessimistic sentiments and generates slogans that resonate emotionally with voters. The key objectives of this study are twofold: first, to accurately classify election campaign slogans into two primary sentiment categories, optimistic and pessimistic, and second, to generate emotionally resonant slogans that can effectively connect with voters. Extensive experiments and sentiment analyses are conducted using a diverse dataset of election campaign slogans to assess the efficiency of CRFVReC. The results highlight the model's remarkable precision in sentiment classification, demonstrating its capability to discriminate between optimistic and pessimistic sentiment in slogans. The model exhibits elevated accuracy, precision, recall, and AUC scores in sentiment classification, utilizing a diverse dataset. Furthermore, CRFVReC showcases its creative potential in generating slogans with emotionally compelling content. Its capability holds significant promise for campaign strategists and political communicators seeking to craft slogans that resonate with voters deeply emotionally. Additionally, the model's adaptability to slogans of varying lengths makes it a versatile tool for election campaign management and strategy development. The CRFVReC emerges as a robust and adaptable solution for sentiment-based slogan generation and analysis in the complex landscape of election campaigns. Its contributions lie not only in inaccurate sentiment classification but also in its potential to shape the narrative of political campaigns through the creation of emotionally impactful slogans. This research contributes to the fields of political communication and campaign management, providing valuable tools and insights for practitioners and researchers.

This paper presents a hybrid image storage model for big data environments. The model combines relational and non-relational (NoSQL) databases, file systems (IPFS), and blockchain technologies to ensure an optimal balance between performance, scalability, and security in image storage. The existing approaches to organising image data storage and image compression methods in decentralised systems are analysed. Optimised image indexing is proposed to accelerate data search and access. A prototype system based on the proposed model was developed, and an experimental study was conducted on various image datasets (medical, satellite, and digital art). The experimental results demonstrate that the hybrid model outperforms traditional approaches: image access time is reduced by ~30% compared to standalone storage systems, providing high scalability (with increased nodes, processing time decreases nonlinearly). The efficiency of image compression in reducing storage costs in blockchain-oriented systems is also confirmed: the WebP format allows file size to be reduced by 40–60% while maintaining acceptable quality (PSNR > 30 dB). The proposed solution is relevant for medical diagnostics, video surveillance systems, geographic information systems, and other fields requiring reliable storage and fast processing of large-scale image datasets.

Accurate liver and tumor segmentation from medical imaging plays an important role in effective diagnosis and appropriate treatment planning, especially in the case of liver cancer. This research proposed a novel U-Net architecture enhanced with image colorization techniques for precise liver tumor segmentation in clinical CT images. The proposed image colorization-based U-Net, which integrates both grayscale-based and RGB-based architectures, was tested on the LiTS dataset and real clinical data. This evaluation aimed to measure its effectiveness in liver and tumor segmentation across different imaging conditions. The grayscale-based U-Net achieved high segmentation accuracy, achieving a DICE coefficient of 99.95% for liver segmentation and 90.44% for tumor segmentation. This strong performance suggests its ability to precisely delineate anatomical structures. The model also achieved an RMSE of 0.019, a PSNR of 82.14, and a pixel accuracy of 0.316, reflecting its capability to reduce reconstruction while preserving overall image quality. These findings further support the model’s reliability in challenging imaging scenarios, suggesting its potential as an effective tool for liver tumor segmentation. To further validate its real-world applicability, the model was tested on clinical data, where it effectively segmented liver and tumor regions across diverse imaging conditions. By addressing challenges such as low contrast and variability in tumor characteristics, the use of grayscale-based colorization techniques enhances feature representation, leading to improved segmentation outcomes. The findings demonstrate the potential of the proposed approach to enhance liver and tumor localization, providing a robust framework for clinical applications.

Stock price prediction anticipates future stock prices using historical data and computational models to assist and guide investing decisions. In financial forecasting, accuracy and efficacy in stock price prediction are essential for making better choices. This research describes a hybrid deep learning strategy for improving the extraction and interpretation of the crucial details from stock price time series data. Traditional approaches confront challenges such as computational complexity and nonlinear stock prices. The suggested method pre-processes stock data with Moving Average Z-Transformation, which emphasises long-term trends and reduces fluctuations in the short term. It combines a Transformed Moving Average Fast-RNN Hybrid with Advanced CNNs to create an efficient computational framework. The Enhanced Deep-CNN layer comprises convolutional layers, batch normalisation, leaky ReLU activations, dropout, max pooling and a dense layer. The performance of the model is quantified using metrics including Mean Absolute Error (MAE), Mean Square Error (MSE), Root Mean Square Error (RMSE), and R-squared (R2). It shows superior prediction accuracy with MAEs of 0.28, 0.15, 0.34, 0.17, and 0.13 for Kotak, ICICI, Axis, and SBI, respectively, outperforming previous models. These measurements provide detailed information about the model's predictive skills, proving its ability to improve stock price forecast accuracy significantly.

This work is devoted to developing a novel transfer learning approach for solving binary semantic segmentation problems that often arise on short samples in the medical (segmentation of nodules in lungs, tumors, polyps, etc.) and other domains. The goal is to optimally select the most suitable dataset from a different subject area with similar feature space and distribution to the target data. Examples show that a severe disadvantage of transfer learning is the difficulty of selecting an initial training sample for pre-training a neural network. In this paper, we propose metrics for calculating the distance between binary segmentation datasets, allowing us to select the optimal initial training set for transfer learning. These metrics are based on the geometric distances estimation of the dataset using optimal transport, Wasserstein distance for Gaussian mixture models, clustering, and their hybrids. Experiments on datasets of medical segmentation Decathlon, LIDC, and a private dataset of tuberculomas in the lungs are presented, proving a statistically strict correlation of these metrics with a relative increase in segmentation accuracy during transfer learning.

Data security has become a major concern in the present era of the communication revolution, especially maintaining the confidentiality of medical images a prime concern in e-health establishments. As conventional techniques hold numerous drawbacks, this study aims to develop an image encryption algorithm by combining two renowned methods: the RSA algorithm and steganography. The proposed algorithm is modified with the help of the conventional RSA algorithm and steganography to provide an attainable solution to this alarming issue. RSA technique encrypts multiple medical images with distinct keys; further, these keys are embedded in two images to be transmitted secretly with the help of LSB steganography. The proposed system generates images of an unidentifiable pattern after encryption and decrypts those images without any loss. The claimed performances and robustness of the system are justified using different numerical and graphical measures such as PSNR, MSE, SSIM, NPCR, UACI, and histograms. This encryption method can be used for medical image transmission where image security is a vital concern.