FaceWarehouse

FaceWarehouse是用于视觉计算应用的3D面部表情数据库。使用Kinect RGBD相机，我们捕获了来自不同种族背景的7-80岁的150个人。对于每个人，我们捕获了她不同表情的RGBD数据，包括中性表情和19种其他表情，例如张开嘴，微笑，亲吻等。对于每个RGBD原始数据记录，彩色图像上的一组面部特征点 (例如眼角，嘴轮廓和鼻尖) 会自动定位，如果需要更好的精度，则会手动调整。然后，我们对模板面部网格进行变形，以尽可能接近地拟合深度数据，同时将彩色图像上的特征点与其在网格上的相应点进行匹配。从这些拟合的脸部网格中，我们为每个人构建了一组特定于个人的表情融合形状。

FaceWarehouse is a 3D facial expression database for visual computing applications. Using a Kinect RGBD camera, we captured RGBD data from 150 individuals aged 7 to 80 years old across diverse ethnic backgrounds. For each individual, we captured RGBD data of various expressions, including the neutral expression and 19 other expressions such as mouth opening, smiling, kissing, and so on. For each raw RGBD recording, a set of facial landmark points (e.g., eye corners, mouth contours and the tip of the nose) on the color image are automatically localized, and manually adjusted if higher accuracy is required. We then deformed a template facial mesh to fit the depth data as closely as possible, while matching the landmark points on the color image to their corresponding points on the mesh. From these fitted facial meshes, we constructed a set of person-specific expressive blendshapes for each individual.

搜集汇总

背景与挑战

背景概述

FaceWarehouse是一个用于视觉计算应用的3D面部表情数据库，包含150名不同种族背景、年龄在7-80岁人群的RGBD数据，每人采集了中性表情和19种其他表情（如张开嘴、微笑等）。数据集通过自动定位和手动调整面部特征点，并变形模板网格来拟合深度数据，为每个人构建了特定于个人的表情融合形状，适用于3D面部建模和表情分析研究。