Walaa M. Abd-Elhafiez

Recent advances in generative modeling have enabled the creation of highly realistic deepfake facial images, posing significant risks to digital security, media integrity, and public trust. Although deep learning–based detection methods have achieved strong performance, they often suffer from limited cross-dataset generalization, sensitivity to manipulation-specific artifacts, and reduced interpretability. To address these limitations, this paper proposes a forensic-first hybrid deepfake face detection framework that integrates handcrafted local forensic descriptors with multi-CNN deep semantic representations. Specifically, manipulation-sensitive regions are captured using a Bag-of-Visual-Words (BoVW) model constructed from Histogram of Oriented Gradients (HOG) features extracted at salient keypoints detected via SURF, FAST, and BRISK. In parallel, high-level features are obtained from fine-tuned ResNet-50, MobileNet, and ShuffleNet models and fused at the feature level to capture complementary semantic information. The combined feature representation is classified using a Support Vector Machine (SVM), enabling stable decision boundaries and improved generalization. Extensive experiments on six benchmark datasets of varying scale and complexity demonstrate that the proposed approach consistently outperforms state-of-the-art methods, achieving up to 97.55% accuracy while maintaining robustness under cross-dataset and challenging forensic conditions. The results highlight the effectiveness of integrating explicit forensic features with deep representations to achieve a robust, interpretable, and generalizable solution for deepfake face detection.

Coffee, the world’s most traded tropical crop, is vital to the economies of many producing countries. However, coffee leaf diseases pose a serious threat to coffee quality and sustainable production. Deep learning has shown strong performance in plant disease identification through automatic image classification. Nevertheless, reliance on a single convolutional neural networks (CNNs) architecture restricts feature variability and real-world generalization. Moreover, limited work has systematically combined feature selection/reduction with CNNs, which constrains the advancement of hybrid models capable of capturing complementary features while ensuring computational efficiency without accuracy loss. This article presents an enhanced deep learning-based framework for coffee disease classification incorporating a hybrid strategy that integrates CNNs and advanced feature selection algorithms. GoogLeNet and ResNet18 are paired for complementary feature extraction, Principal Component Analysis (PCA) and Singular Value Decomposition (SVD) are employed for dimensionality reduction, and ANOVA and Chi-square are applied to select the most informative features. An Adam optimizer (learning rate = 0.001, batch size = 20, epochs = 50) with early stopping is used for training. Experiments on the BRACOL dataset achieved 99.78% accuracy, with precision, recall, and F1-score all exceeding 99% across classes. To the best of our knowledge, this study systematically integrates GoogLeNet and ResNet18 with PCA/SVD dimensionality reduction and analysis of variance (ANOVA)/Chi-square feature selection, for coffee disease classification, thereby addressing a key gap in prior research.

Egypt’s geographic location and natural resources make it an ideal candidate for Renewable Energy (RE) projects. The government has implemented initiatives and policies to promote RE development, resulting in notable projects such as the Benban Solar Park and wind farms in the Gulf of Suez. These endeavors aim to combat climate change, reduce greenhouse gas emissions, and foster sustainable development. Multicriteria Decision Making (MCDM) techniques are increasingly favored due to their ability to accommodate diverse criteria for selecting RE sources. This paper evaluates numerous RE options in Egypt, including solar, wind, hydropower, and biomass energy, and proposes a hybrid strategy for classifying RE sources for power generation based on criteria such as finance, environment, technology, and community. The study utilizes four MCDM methods: the Analytical Hierarchy Process (AHP), the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS), Fuzzy TOPSIS (FTOPSIS), and Fuzzy AHP (FAHP). The key findings indicate that solar and wind energy are the most viable options for Egypt’s renewable energy strategy, with the FTOPSIS method proving to be the most effective for ranking these sources by providing a comprehensive evaluation framework that aligns well with related studies. These findings have significant implications for policy and investment, as they provide a robust decision-making framework that ensures balanced consideration of financial, environmental, technological, and social factors. The study underscores the value of using advanced MCDM techniques to guide RE development in Egypt, contributing to the global effort to combat climate change.

In this paper, we present a new encryption method based on discrete wavelet transform (DWT). This method provides a number of advantages as a pseudo randomness and sensitivity due to the variation of the initial values. We start by decomposing the image with spatial reconstruction by DWT, followed by preformation by fractional chaotic cryptovirology and Henon map keys for space encryption. Bearing in mind the rapid pace of change in the fields of fractional chaotic cryptovirology, our findings suggest that the proposed method is too powerful relative to the key sensitivity, histogram, entropy and correlation coefficient.

In this paper, we combine decision fusion methods with four meta-heuristic algorithms (Particle Swarm Optimization (PSO) algorithm, Cuckoo search algorithm, modification of Cuckoo Search (CS McCulloch) algorithm and Genetic algorithm) in order to improve the image segmentation. The proposed technique based on fusing the data from Particle Swarm Optimization (PSO), Cuckoo search, modification of Cuckoo Search (CS McCulloch) and Genetic algorithms are obtained for improving magnetic resonance images (MRIs) segmentation. Four algorithms are used to compute the accuracy of each method while the outputs are passed to fusion methods. In order to obtain parts of the points that determine similar membership values, we apply the different rules of incorporation for these groups. The proposed approach is applied to challenging applications: MRI images, gray matter/white matter of brain segmentations and original black/white images Behavior of the proposed algorithm is provided by applying to different medical images. It is shown that the proposed method gives accurate results; due to the decision fusion produces the greatest improvement in classification accuracy.

A new method to enhance image encryption is proposed. It is based on lifting transform and edge detection technique, the edge detection is selected as a way that it increases the capacity of hidden data, which smooth (non-edge) area hide small amount of information as compared to edge area. Initially, the lifting transform will apply on entering image. For approximate part, the simplest edge detection technique is applied, then, the image into N×N blocks is divided, the selection of key for edge will be different about non-edge block. To make the proposed method better enough, different techniques are used, one for the edge block and another for non-edge block. The proposed method is proved resists in most of security attacks. Further different testes analysis as entropy, correlation and histogram analysis is used to verify the performance of proposed method.

In this paper, two image encryption schemes are proposed for grayscale and color images. The two encryption schemes are based on dividing each image into blocks of different sizes. In the first scheme, the two dimension (2D) input image is divided into various blocks of size N × N. Each plainimage block is transformed into a one dimensional (1D) array using the Zigzag pattern mode. Then, the exclusive or (XOR) logical operation is used to encrypt each block with the analogous secret key. In the second scheme, after the transformation process, the first block of each image is encrypted by the corresponding secret key. Then, before the next block is encrypted, it is XORed with the first encrypted block to become the next input to the encrypting routine and so on. This feedback mechanism depends on the cipher block chaining (CBC) mode of operation which considers the heart of some ciphers because it is highly nonlinear. In the case of color images, the color component is separated into blocks with the same size and different secret keys. The used secret key sequences are generated from elliptic curves (EC) over a binary finite field F ₂ ™. Finally, the experimental results are carried out and security analysis of the ciphered images are demonstrated that the two proposed schemes had a better performance in terms of security, sensitivity and robustness.

This paper provides a simple and efficient method to detect human faces in videos and recognize persons within the video according to a preset database of known persons. The proposed system consists of three main steps. The first step is skin-like regions detection in CIE-Luv color space. The second step is face detection based on skin-like regions, contour detection and geometrical properties such as face shape. The third step is face verification, in which, each face is compared with a gallery of known faces and the location of the best matched one is returned. Face verification step uses variance formula and skin to non-skin percentage in each facial feature to compare the test face and the faces images in the known database. The proposed system was tested on many different videos with different number of persons in the video. The detected faces are compared with a preset database of known images. Experimental results show that the proposed system is efficient enough to detect faces in different lighting conditions, head pose and face expressions. The results of verification step show that the proposed system is able to retrieve faces from a database with good accuracy in a reasonable computational time compared with classical method, variance estimation method and facial extraction method.