OREBA

OREBA数据集是由纽卡斯尔大学创建的，旨在通过多传感器数据记录共餐过程中的饮食行为和相关摄入，以支持自动检测摄入手势的研究。该数据集包含两种情景：单份餐和共享餐，总计202次餐记录，涉及180名参与者，记录了9069次摄入手势。数据集提供了同步的前视视频和IMU（包括加速度计和陀螺仪）数据，以及每次摄入手势的标签。OREBA数据集的应用领域主要集中在自动饮食监测，特别是通过机器学习方法检测和分析共餐环境中的饮食行为，以解决传统自我报告数据方法的局限性。

The OREBA dataset was developed by Newcastle University, aiming to record dietary behaviors and associated intake during shared meals via multi-sensor data to support research on automatic eating gesture detection. This dataset includes two scenarios: single-serving meals and shared meals, with a total of 202 meal recordings involving 180 participants and 9069 recorded eating gestures. The dataset provides synchronized front-facing video and IMU (including accelerometer and gyroscope) data, as well as labels for each eating gesture. The primary application domains of the OREBA dataset focus on automatic dietary monitoring, particularly detecting and analyzing dietary behaviors in shared meal environments using machine learning methods, to address the limitations of traditional self-reported data approaches.