Artificial Intelligence
This repository contains the datasets produced using different data generation strategies to train data driven models (e.g., decision trees, gradient tree boosting, and deep neural networks), and to evaluate their performances. The data generation strategies are described, and the results are presented in the conference paper: "Training Data Generation Strategies for Data-driven Security Assessment of Low Voltage Smart Grids" J. Cuenca, E. Aldea, E. Le Guern-Dall'o, R. Féraud, G. Camilleri, and A. Blavette. IEEE ISGT EU 2024, Dubrovnik, Croatia, Oct 2024.
- Categories:
44 Views
Human pose estimation has applications in numerous fields, including action recognition, human-robot interaction, motion capture, augmented reality, sports analytics, and healthcare. Many datasets and deep learning models are available for human pose estimation within the visible domain. However, challenges such as poor lighting and privacy issues persist. These challenges can be addressed using thermal cameras; nonetheless, only a few annotated thermal human pose datasets are available for training deep learning-based human pose estimation models.
- Categories:
306 Views
This dataset is a sequential recommendation dataset that includes three sub-datasets: Beauty, Toys, and Yelp, specifically designed for research and development in recommendation systems. All datasets have been pre-processed, allowing users to directly input them into the main program for use. These datasets are ready for experiments involving user-item interactions and can be used to train and evaluate recommendation algorithms. The command to run the datasets is: .
- Categories:
56 Views
The Multi-Server Multi-User computation offloading dataset is a dataset based on the scenario of multi-server multi-user binary computing offloading. It is characterized by the connection status between users and edge servers, user task information, and server computational resource information. The solution aims to minimize the total cost of power consumption and latency of all tasks. The labels are the offloading decisions of user tasks and the computational resource allocation of edge servers. The features and labels of this dataset are graph-structured.
- Categories:
142 Views
Acoustic non-line-of-sight vehicle detection dataset. Complete multi-channel audio of vehicles passing through the intersection was captured at multiple intersections. It can be used for acoustic non-line-of-sight vehicle detection. The direction in which the vehicle entered the intersection, and the moments when the vehicle entered and left the line of sight were recorded in the file names, and the audio categories can be classified based on that moment. All audio files were stored in five folders to facilitate five-fold cross-validation.
- Categories:
37 Views
Acoustic non-line-of-sight vehicle detection dataset. Complete multi-channel audio of vehicles passing through the intersection was captured at multiple intersections. It can be used for acoustic non-line-of-sight vehicle detection. The direction in which the vehicle entered the intersection, and the moments when the vehicle entered and left the line of sight were recorded in the file names, and the audio categories can be classified based on that moment. All audio files were stored in five folders to facilitate five-fold cross-validation.
- Categories:
15 Views
The data comes from V2X-Sim full dataset. We keep vehicle onboard camera data, LiDAR data and infrastructure data in V2X-Sim. It encompasses a comprehensive and well-
synchronized collection of both infrastructure and vehicle sensor data. It also contains well-annotated ground truth data,which can support multiple perception tasks such as
detection, segmentation, and tracking. It includes six views of vehicle-side images and four views of infrastructure-side images. It collects 100 scenes in total, and every
scene includes 100 records.
- Categories:
27 Views
This paper presents an innovative Internet of Things (IoT) system that integrates gas sensors and a custom Convolutional Neural Network (CNN) to classify the freshness and species of beef and mutton in real time. The CNN, trained on 9,928 images, achieved 99% accuracy, outperforming models like ResNet-50, SVM, and KNN. The system uses three gas sensors (MQ135, MQ4, MQ136) to detect gases such as ammonia, methane, and hydrogen sulfide, which indicate meat spoilage.
- Categories:
292 Views
To provide a standardized approach for testing and benchmarking secure evaluation of transformer-based models, we developed the iDASH24 Homomorphic Encryption track dataset. This dataset is centered on protein sequence classification as the benchmark task. It includes a neural network model with a transformer architecture and a sample dataset, both used to build and evaluate secure evaluation strategies.
- Categories:
130 Views
The quality and safety of tea food production is of paramount importance. In traditional processing techniques, there is a risk of small foreign objects being mixed into Pu-erh sun-dried green tea, which directly affects the quality and safety of the food. To rapidly detect and accurately identify these small foreign objects in Pu-erh sun-dried green tea, this study proposes an improved YOLOv8 network model for foreign object detection. The method employs an MPDIoU optimized loss function to enhance target detection performance, thereby increasing the model's precision in targeting.
- Categories:
68 Views