There are a variety of potential uses for the classification of eye conditions, including tiredness detection, psychological condition evaluation, etc. Because of its significance, many studies utilizing typical neural network algorithms have already been published in the literature, with good results. Convolutional neural networks (CNNs) are employed in real-time applications to achieve two goals: high accuracy and speed. However, identifying drowsiness at an early stage significantly improves the chances of being saved from accidents. Drowsiness detection can be automated by using the potential of artificial intelligence (AI), which allows us to assess more cases in less time and with a lower cost. With the help of modern deep learning (DL) and digital image processing (DIP) techniques, in this paper, we suggest a CNN model for eye state categorization, and we tested it on three CNN models (VGG16, VGG19, and 4D). A novel CNN model named the 4D model was designed to detect drowsiness based on eye state. The MRL Eye dataset was used to train the model. When trained with training samples from the same dataset, the 4D model performed very well (around 97.53% accuracy for predicting the eye state in the test dataset). The 4D model outperformed the performance of two other pretrained models (VGG16, VGG19). This paper explains how to create a complete drowsiness detection system that predicts the state of a driver’s eyes to further determine the driver’s drowsy state and alerts the driver before any severe threats to road safety.

Cricket shots cannot be detected yet from single video sample without multiple view camera and other tools like sonar, speedometer. Extracting salient feature and optical flow from videos of cricket shots is still a challenge. In cricket, body parts movement created several different directional optical flows. So we propose motion estimation approach related to classifying the shots using 3D MACH for action recognition. Our methodology defines 8 classes of angle ranges to detect cricket shots. Our method is grounded on Motion vectors that help to measure the angle of any precise cricket shot. An adequate accuracy level for the shots is established for this particular approach.

A textual data processing task that involves the automatic extraction of relevant and salient keyphrases from a document that expresses all the important concepts of the document is called keyphrase extraction. Due to technological advancements, the amount of textual information on the Internet is rapidly increasing as a lot of textual information is processed online in various domains such as offices, news portals, or for research purposes. Given the exponential increase of news articles on the Internet, manually searching for similar news articles by reading the entire news content that matches the user’s interests has become a time-consuming and tedious task. Therefore, automatically finding similar news articles can be a significant task in text processing. In this context, keyphrase extraction algorithms can extract information from news articles. However, selecting the most appropriate algorithm is also a problem. Therefore, this study analyzes various supervised and unsupervised keyphrase extraction algorithms, namely KEA, KP-Miner, YAKE, MultipartiteRank, TopicRank, and TeKET, which are used to extract keyphrases from news articles. The extracted keyphrases are used to compute lexical and semantic similarity to find similar news articles. The lexical similarity is calculated using the Cosine and Jaccard similarity techniques. In addition, semantic similarity is calculated using a word embedding technique called Word2Vec in combination with the Cosine similarity measure. The experimental results show that the KP-Miner keyphrase extraction algorithm, together with the Cosine similarity calculation using Word2Vec (Cosine-Word2Vec), outperforms the other combinations of keyphrase extraction algorithms and similarity calculation techniques to find similar news articles. The similar articles identified using KPMiner and the Cosine similarity measure with Word2Vec appear to be relevant to a particular news article and thus show satisfactory performance with a Normalized Discounted Cumulative Gain (NDCG) value of 0.97. This study proposes a method for finding similar news articles that can be used in conjunction with other methods already in use.

Healthcare data is highly sensitive and must be safeguarded. Personal and sensitive data, such as names and addresses, is stored in Encrypted Electronic Health Records (EHRs). This paper proposes a Blockchain-based distributed application platform for Bangladesh’s public and private healthcare service providers. The proposed application framework enables users to create secure digital agreements for commerce or collaboration by leveraging data immutability and smart contracts. As a result, all stakeholders can collaborate securely over the same Blockchain network, taking advantage of their data’s openness and read/write nature. The proposed application is made up of various application interfaces for various stakeholders. The proposed solution employs Hyperledger Fabric and Blockchain to ensure data integrity, privacy, permissions, and service availability. In the application portal, each user has a profile. The creation of a unique identity for each user, as well as the establishment of digital information centers across the country, has greatly aided the process. This application collects health data from each user in a systematic manner, which is useful for research institutes and healthcare-related organizations. For this application, a national data warehouse in Bangladesh is feasible, and various healthcare-related analyses can be performed using the collected data, assisting the strategy and planning department in making informed decisions regarding the healthcare sector in Bangladesh. Because Bangladesh has both public and private healthcare providers, a simple digital strategy is essential for all organizations to accomplish their services. This study proposes a solution to achieve this goal.

This study presents FireNet-CNN, an advanced deep-learning model particularly designed for forest fire detection, which significantly surpasses existing methods in terms of reliability, efficiency, and interpretability. FireNet-CNN is compared to popular pre-trained models, including VGG16, VGG19, and Inception V3, across key performance metrics and consistently shows superior results, achieving 99.05% accuracy, 99.41% precision, and 98.28% recall. The model was evaluated using two augmented datasets: Dataset A and Dataset B, which consist of fire and non-fire images sourced from multiple video and image datasets. FireNet-CNN’s architecture, which includes 2.75 million parameters and a compact model size of 10.58 MB, has been meticulously optimized for fire detection tasks. As a consequence, the inference time of 0.95 seconds/image enables fast real-time deployment especially suitable for resource-constrained platforms like drones, remote sensors or other types of embedded systems in wooded regions. FireNet-CNN uses synthetic data augmentation based on Stable Diffusion to overcome the limitations of dataset size and class imbalance. This augmentation is critical as it helps the model accurately identify fire instances with a lower false positive rate, which is key for any real-time fire detection system where reliability and dependability are vital. To improve transparency and trust in safety-critical applications, FireNet-CNN incorporates the explainable AI (XAI) techniques, such as Grad-CAM and Saliency Maps. Despite encountering challenges such as reliance on synthetic data and issues of class imbalance, FireNet-CNN has demonstrated promising potential as a viable and effective solution for early wildfire detection. It offers significant insights for future research and practical applications in fire management and disaster response.

Human detection is a special application of object recognition and is considered one of the greatest challenges in computer vision. It is the starting point of a number of applications, including public safety and security surveillance around the world. Human detection technologies have advanced significantly in recent years due to the rapid development of deep learning techniques. Despite recent advances, we still need to adopt the best network-design practices that enable compact sizes, deep designs, and fast training times while maintaining high accuracies. In this article, we propose ReSTiNet, a novel compressed convolutional neural network that addresses the issues of size, detection speed, and accuracy. Following SqueezeNet, ReSTiNet adopts the fire modules by examining the number of fire modules and their placement within the model to reduce the number of parameters and thus the model size. The residual connections within the fire modules in ReSTiNet are interpolated and finely constructed to improve feature propagation and ensure the largest possible information flow in the model, with the goal of further improving the proposed ReSTiNet in terms of detection speed and accuracy. The proposed algorithm downsizes the previously popular Tiny-YOLO model and improves the following features: (1) faster detection speed; (2) compact model size; (3) solving the overfitting problems; and (4) superior performance than other lightweight models such as MobileNet and SqueezeNet in terms of mAP. The proposed model was trained and tested using MS COCO and Pascal VOC datasets. The resulting ReSTiNet model is 10.7 MB in size (almost five times smaller than Tiny-YOLO), but it achieves an mAP of 63.74% on PASCAL VOC and 27.3% on MS COCO datasets using Tesla k80 GPU.