he Human Activity Recognition (HAR) dataset consists of over 12,000 labeled images spanning 15 distinct human activity classes, including actions such as calling, dancing, running, and sleeping. Each image represents a single activity and is organized into separate folders corresponding to its class label. The objective is to develop a convolutional neural network (CNN)-based image classification model capable of accurately predicting the activity being performed in each image. A separate test set of 5,400 unlabeled images is provided for evaluation, along with a submission template that specifies the required format for output predictions. This task falls under the broader domain of computer vision and has practical applications in surveillance, healthcare monitoring, human-computer interaction, and behavior analysis.

This dataset contains a vast collection of 26.9k images, which have been carefully annotated for the specific purpose of ship detection. The bounding box annotations are presented in the YOLO format, which allows for accurate and efficient detection of the ships in the images. The dataset has been curated to include images of only one class - "ship" - thus enabling streamlined and precise analysis.

The detection of ships or vessels within an image is a vital task that has significant practical applications. Maritime safety is one such application, as the detection of ships can help prevent accidents at sea by providing early warnings of potential collisions or obstacles. Fisheries management is another important use case, where the detection of fishing vessels can aid in monitoring fishing activities and preventing overfishing. In addition, ship detection can be used for marine pollution monitoring, defense and maritime security, protection from piracy, illegal immigration, and a range of other purposes.

The ABC-XYZ Inventory Classification Dataset is a synthetic dataset designed to help supply chain professionals and data enthusiasts practice inventory segmentation through ABC and XYZ classification techniques. It features 1,000 unique items across five categories, with 12 months of demand data, unit prices, and annual sales values. The dataset simulates real-world demand variability and includes built-in patterns for value-based prioritization (ABC) and demand stability analysis (XYZ), making it ideal for inventory optimization and cost-reduction strategies.

The Face Mask Detection Dataset includes 853 labeled images categorized into three classes: people wearing masks, not wearing masks, and wearing masks incorrectly. Each image comes with bounding box annotations in PASCAL VOC format. This dataset enables the development of deep learning models to detect mask usage and assess its correctness, offering a valuable tool for public health applications, especially during pandemics like COVID-19.