BlanketGen2-Fit3D
收藏资源简介:
BlanketGen2-Fit3D是由葡萄牙科学技术基金会支持创建的数据集,旨在通过合成逼真的毯子遮挡数据来改进床上的2D人体姿态估计。该数据集基于Fit3D数据集,包含1,217,312帧图像,通过BlanketGen2管道生成,利用SMPL模型和布料模拟技术合成毯子遮挡效果。数据集的应用领域主要集中在医疗监控中的床上姿态估计,旨在解决毯子遮挡下的人体姿态识别问题。
BlanketGen2-Fit3D is a dataset developed with funding support from the Portuguese Foundation for Science and Technology. It is created to enhance 2D human pose estimation for in-bed environments by generating realistic synthetic blanket occlusion data. Built on the Fit3D dataset, this collection comprises 1,217,312 image frames, all generated via the BlanketGen2 pipeline, which leverages the SMPL model and cloth simulation technologies to synthesize blanket occlusion effects. Its primary application areas focus on in-bed pose estimation for medical monitoring, with the goal of solving the challenge of human pose recognition under blanket occlusion.




