Pixel-Face
收藏资源简介:
Pixel-Face是由阿里巴巴集团等机构创建的大型高分辨率3D人脸数据集,包含855个年龄在18至80岁之间的个体,每个个体拥有超过20种不同表情的样本。数据集内容丰富,包括高分辨率的多视角RGB图像和3D网格,以及精确的面部标志注释和3D注册结果。创建过程中,使用了高精度的三目结构光系统进行数据采集,并通过多视角融合和注册确保数据质量。Pixel-Face的应用领域广泛,主要用于推动3D人脸重建技术的研究,解决现有数据集在数量、真实性和多样性上的不足。
Pixel-Face is a large-scale high-resolution 3D human face dataset developed by Alibaba Group and other institutions. It includes 855 individuals aged between 18 and 80 years old, with each subject having samples of over 20 different facial expressions. The dataset features rich content, such as high-resolution multi-view RGB images, 3D meshes, precise facial landmark annotations, and 3D registration results. During the dataset construction, a high-precision trinocular structured light system was utilized for data acquisition, and multi-view fusion and registration techniques were adopted to guarantee data quality. Pixel-Face has broad application prospects, primarily used to advance research on 3D face reconstruction technology by addressing the shortcomings of existing datasets in terms of quantity, authenticity, and diversity.

- 1Pixel-Face: A Large-Scale, High-Resolution Benchmark for 3D Face Reconstruction阿里巴巴集团 · 2020年



