The rise of Deepfake technology in the information age provided significant challenges in ensuring the reliability and authenticity of digital media content. Deepfake pictures, videos, and audio recordings have reached advanced levels of sophistication, making it challenging to distinguish authenticity from fraudulent activity. Deepfakes are highly manipulated images, audio recordings, and videos that use artificial intelligence to create convincing forgeries of individuals engaging in actions or making statements they never actually performed. This advanced technology has multiple applications, including entertainment and social media, as well as potentially harmful activities like propagandism and disinformation. Innovative AI techniques, such as deep learning, transfer learning, Long Short-Term Memory (LSTM) networks, and Convolutional Neural Networks (CNNs), have been developed to effectively detect and combat Deepfakes. Deep learning algorithms, specifically CNNs, which are inspired by the visual cortex of the human brain, have proven to be effective in detecting Deepfake images and videos. The solution to this problem is to use collective techniques that improve detection accuracy by combining multiple AI systems. Addressing this challenge involves employing a combination of techniques to enhance detection accuracy, such as leveraging the strengths of CNNs and LSTM models in addition to other techniques like Generative Adversarial Networks (GANs). The primary goal is to provide studies on protecting the integrity and authenticity of digital content using different algorithms and standard datasets, while also investigating potential future developments in the field. AI techniques effectively detect Deepfake in various media, improving digital content authenticity. Deepfake technology’s growth presents risks in fraud and misinformation, urging advancements in detection methods. The study advocates for collaborative efforts and legal reforms, especially in India, to combat Deepfake challenges.

To digitize and preserve the cultural heritage in the form of Indian classical dance become apparent area of research. Sattriya classical dance of North-East India (Assam) is one of the eight Indian classical dance forms that requires immediate preservation. Sattriya classical dance consists of 29 Asamyukta hastas (single-hand gestures) and 14 Samyukta hastas (double-hand gestures). Moreover, the foundation of Samyukta hasta depends on understanding Asamyukta hasta. Therefore, the paper aims to classify single-hand gestures of Sattriya classical dance only. Although, a solution based on two level classification method to classify the Sattriya classical dance is available in recent literature, but it requires a trial and error method to select the optimized features. Since, Asamyukta hastas can appear closely similar to each other and therefore misclassification chances are very high. Thus, accuracy rate obtained for the two level classification method was only 75.45%. So, to address this issues in this paper, a Multilevel Classification Model with Vision based Features (MCM-\(V_b\)F) has been proposed to classify the Asamyukta hastas of Sattriya classical dance. This model uses two types of feature matching, viz., high-level feature matching and low-level feature matching. To extract the high-level features and low-level features different algorithm has been proposed. In this model, features are automatically selected. This proposed MCM-\(V_b\)F model is also tested on Asamyukta hasta mudras of Bharatanatyam classical dance of South India (Tamil Nadu). This model obtain an accuracy 94.12%, 87.14% for Sattriya classical dance Single-Hand Gestures (SSHG) dataset and Bharatnatyam classical dance Single-Hand Gestures (BHSG) dataset respectively. This paper also provides the comparative study of the proposed model MCM-\(V_b\)F with traditional bench-mark classifier model such as Naive Bayes, Decision Tree and Support Vector Classifier (SVM) etc.

This paper presents an automated helmet detection system leveraging Raspberry Pi and advanced deep learning techniques, specifically the YOLOv9 object detection model that focuses on real-time object detection. The system captures real-time video frames of motorcyclists, processes these frames to detect helmet usage, and interfaces with the vehicle’s ignition control unit to inhibit engine start if a helmet is not detected. This edge computing solution addresses the inefficiencies of manual inspections by providing a scalable, cost-effective, and real-time enforcement mechanism for helmet compliance. The YOLOv9 model, trained on a robust dataset, ensures high detection accuracy and low latency, demonstrating effective performance in real-world scenarios. The YOLOv9 model achieved a precision of 0.894, recall of 0.943, mAP50 of 0.967, and an inference speed of 21.0 ms per image. This research highlights significant advancements in intelligent transportation systems and road safety, achieved through integrating convolutional neural networks and embedded hardware platforms.

Nail diseases pose significant health concerns and often require prompt diagnosis and treatment. The authors propose a new approach for identifying nail diseases using advanced deep learning (DL) techniques. Specifically, we employ a modified DenseNet169 architecture, integrating Leaky Rectified Linear Unit (ReLU) activation and Long Short-Term Memory (LSTM) layers to extract features from nail images effectively. Our methodology involves pre-processing the images, training the modified DenseNet169-LSTM model, data balancing, and evaluating its performance using various metrics. The proposed method achieved an F1 score of 89.9%, while average Area Under the Curve of 98.2%, F1 score of 84.7%, Matthews correlation coefficient (MCC) of 84.7% and a Kappa score of 84.6%, with 95% confidence intervals (CI) of 83.7% (lower) and 87.3% (higher) and a p-value of 0.016. Moreover, the method’s robustness was also tested using the 5-fold method. The proposed approach demonstrates promising results in accurately identifying nail diseases, offering potential applications in clinical settings for timely diagnosis and treatment.