Multi-View Cow Image Dataset for AI-Based Weight Estimation

This dataset contains a multi-view collection of RGB images of cows designed for artificial intelligence–based weight estimation. The dataset includes 360 images collected from 72 individual cows, where each cow is photographed from five predefined angles to capture comprehensive body shape information. Images were collected in real outdoor environments in Dhaka, Bangladesh, using smartphone cameras under natural lighting conditions.
The dataset is organized into angle-based folders and follows a standardized file naming convention that encodes cow ID and capture angle. In addition to the image data, the dataset includes a labels file containing cow-level weight annotations and a metadata file providing image-level contextual information such as capture angle, collection date, device used, and collector name.
This dataset supports research in computer vision, machine learning, and precision livestock farming. It is particularly suitable for developing and evaluating image-based regression models for non-invasive cow weight estimation in real-world agricultural settings.