EgoHands

EgoHands是一个识别复杂手部运动的数据集。该数据集包含48个Google Glass视频，这些视频涉及两个人之间的复杂第一人称互动。数据集的主要目的是能够更好地使用数据驱动的方法来理解手的运动，该方法从语义上区分观察者的手和其他人的手，以及区分左右手。演员可以自由参与一系列互动，手的姿势几乎不受约束。

EgoHands is a dataset designed for recognizing complex hand movements. It contains 48 Google Glass videos that record complex first-person interactions between two people. The primary goal of this dataset is to enable better understanding of hand movements through data-driven approaches that semantically distinguish between the observer's hands and others' hands, as well as differentiate between left and right hands. Actors are free to engage in a variety of interactions, with hand poses being almost unconstrained.

搜集汇总

背景与挑战

背景概述

EgoHands是一个专注于复杂手部运动识别的数据集，包含48个第一人称视角的Google Glass视频，记录两人互动场景。该数据集旨在支持数据驱动方法，以语义区分观察者手部与他人手部以及左右手，手部姿势自由度较高，适用于手部运动理解研究。