The field of 3D reconstruction faces persistent challenges in achieving high-quality models, often hindered by issues related to accuracy, completeness, and real-time generation. To address these issues, deep learning has emerged as a promising solution. However, building and training deep learning models from scratch requires substantial computational costs and large datasets. In this context, transfer learning proves advantageous in mitigating these challenges by leveraging pre-trained models on related domains. In this paper, we present a transfer learning-based approach for 3D reconstruction from a single 2D image. Our approach involves fine-tuning pre-trained models using unseen, high-quality datasets to enhance the accuracy, completeness, and robustness of real-world object reconstructions. In order to demonstrate the effectiveness of our approach, we conducted an experimental study that involved retraining the Pix2Vox model using the pascal3D dataset. Through various metrics and a visual comparison, we demonstrate the advancements achieved by our retrained model over the initial pre-trained one. The obtained results demonstrate improvements in accuracy and completeness.
Protecting Internet of things (IoT) devices against cyber attacks is imperative owing to inherent security vulnerabilities. These vulnerabilities can include a spectrum of sophisticated attacks that pose significant damage to both individuals and organizations. Employing robust security measures like intrusion detection systems (IDSs) is essential to solve these problems and protect IoT systems from such attacks. In this context, our proposed IDS model consists on a combination of convolutional neural network (CNN) and long short-term memory (LSTM) deep learning (DL) models. This fusion facilitates the detection and classification of IoT traffic into binary categories, benign and malicious activities by leveraging the spatial feature extraction capabilities of CNN for pattern recognition and the sequential memory retention of LSTM for discerning complex temporal dependencies in achieving enhanced accuracy and efficiency. In assessing the performance of our proposed model, the authors employed the new CICIoT2023 dataset for both training and final testing, while further validating the model’s performance through a conclusive testing phase utilizing the CICIDS2017 dataset. Our proposed model achieves an accuracy rate of 98.42%, accompanied by a minimal loss of 0.0275. False positive rate (FPR) is equally important, reaching 9.17% with an F1score of 98.57%. These results demonstrate the effectiveness of our proposed CNN-LSTM IDS model in fortifying IoT environments against potential cyber threats.
The rapid increase of IoT devices has led to significant security challenges due to their diverse nature and the complexity of potential cyber threats. This study addresses this problem by exploring reinforcement learning (RL) as a dynamic and adaptable solution for IoT security. The research employs RL within the CyberBattleSim simulation platform environment to develop and evaluate defensive strategies against cyber threats. To achieve this, a custom IoT network environment was designed, mimicking a smart house IoT network pattern and utilizing a Q-learning algorithm selected for its potential to optimize security responses in this simulated IoT network setting. The environmental parameters were designed to mimic realistic IoT scenarios, facilitating a thorough investigation of the RL technique’s efficacy. The results indicate that the Q-learning agent demonstrates superiority in identifying and mitigating threats, highlighting the technique’s potential to enhance IoT security. However, further validation in real-world settings is needed to make the study more accurate.
The security of every system hinges on a robust policy that orchestrates controls to safeguard the confidentiality, integrity, and accessibility of information. Implementing such a policy requires meticulous formulation grounded in mathematical and logical precision. In this context, we present a formal modeling and implementation of the Clark-Wilson security method using the FoCaLiZe environment, a workshop equipped with certification capabilities, where programming is intertwined with formal proof. The proposed approach enables the specification of the Clark-Wilson policy constraints and security principles as properties and theorems within FoCaLiZe. Thanks to Zenon, the automatic theorem prover of FoCaLiZe, derived properties and theorems that ensure system safety can be checked and proven.
Osteosarcoma is a type of bone cancer. Its diagnosis refers to anatomopathology by examining tissue samples under the microscope. The goal of this examination is to determine the morphology and behavior of cancer cells in order to determine the presence and extent of osteosarcoma in the patient. This paper focuses on the use of statistical features to classify osteosarcoma disease. The main idea is to use advanced statistical analysis techniques to provide more accurate and effective analysis.By incorporating statistical features extracted from histopathology images, the study aims to develop a robust classification model based on a gradient boosting algorithm that can help accurately detect and classify osteosarcoma cases. The ultimate goal is to help improve diagnostic methods for osteosarcoma, potentially enhancing treatment options and outcomes for patients. The algorithm was applied to histopathological images of osteosarcoma, which consisted of three categories: viable tumors, non-viable tumors, and non-tumor tumors. The model exhibited impressive performance metrics, with an accuracy of 94.54%, precision of 94.33%, specificity of 93.91%, and sensitivity of 94.61%.
