ALPD (Auto-Labeling of Large-Scale Poultry Datasets)
收藏资源简介:
ALPD数据集是由阿肯色大学和乔治亚大学联合创建的大规模家禽数据集,旨在通过半监督学习、主动学习和提示-检测方法实现高效自动标注。该数据集包含1700条数据,主要来源于肉鸡和蛋鸡的视频数据,经过图像转换、过滤、预处理和增强处理。数据集涵盖了多种家禽行为和健康状况的监测,旨在解决家禽养殖中数据标注耗时、成本高的问题。通过该数据集,研究人员能够训练机器学习模型,提升家禽行为检测和健康监测的精度,推动精准畜牧业的发展。
The ALPD dataset is a large-scale poultry dataset jointly created by the University of Arkansas and the University of Georgia. It aims to achieve efficient automatic annotation via semi-supervised learning, active learning, and prompt-detection methods. The dataset contains 1700 data samples, mainly derived from video data of broilers and laying hens, and has undergone image conversion, filtering, preprocessing, and data augmentation. It covers the monitoring of various poultry behaviors and health conditions, and is designed to solve the problems of time-consuming and high-cost data annotation in poultry farming. Through this dataset, researchers can train machine learning models to improve the accuracy of poultry behavior detection and health monitoring, thereby promoting the development of precision livestock farming.

- 1Efficient Auto-Labeling of Large-Scale Poultry Datasets (ALPD) Using Semi-Supervised Models, Active Learning, and Prompt-then-Detect Approach阿肯色大学, 乔治亚大学 · 2025年



