MERL梅尔购物数据集

MERL 购物数据集由 106 个视频组成，每个视频的序列大约为 2 分钟。这些视频来自一个固定的头顶摄像头，向下看人们在杂货店购物。每个视频都包含以下 5 个动作的多个实例："到达货架"（把手伸进货架）、"从货架上收回"（从货架上收回手）、"手放在货架上"（手放在货架上的延长期）、"检查产品"（在手握产品时检查产品）和"检查货架"（在触摸或接触货架时查看货架）。 数据引用： Singh, B., Marks, T.K., Jones, M.J., Tuzel, C.O., Shao, M., "A Multi-Stream Bi-Directional Recurrent Neural Network for Fine-Grained Action Detection", IEEE Conference on Computer Vision and Pattern Recognition (CVPR), DOI: 10.1109/CVPR.2016.216, June 2016, pp. 1961-1970.

MERL Shopping Dataset consists of 106 video clips, each with an approximate duration of 2 minutes. All videos are captured by a fixed overhead camera, capturing people grocery shopping from a top-down perspective. Each video contains multiple instances of the following five actions: "Reach into Shelf" (reaching a hand into the shelf), "Retract Hand from Shelf" (pulling the hand back from the shelf), "Hand on Shelf" (an extended period when the hand rests on the shelf), "Inspect Product" (examining a product while holding it), and "Inspect Shelf" (viewing the shelf while touching or making contact with it). Dataset citation: Singh, B., Marks, T.K., Jones, M.J., Tuzel, C.O., Shao, M., "A Multi-Stream Bi-Directional Recurrent Neural Network for Fine-Grained Action Detection", IEEE Conference on Computer Vision and Pattern Recognition (CVPR), DOI: 10.1109/CVPR.2016.216, June 2016, pp. 1961-1970.