Human Foot Keypoint

Existing human pose datasets contain limited body part types. The MPII dataset annotates ankles, knees, hips, shoulders, elbows, wrists, necks, torsos, and head tops, while COCO also includes some facial keypoints. For both of these datasets, foot annotations are limited to ankle position only. However, graphics applications such as avatar retargeting or 3D human shape reconstruction require foot keypoints such as big toe and heel. Without foot information, these approaches suffer from problems such as the candy wrapper effect, floor penetration, and foot skate. To address these issues, a small subset of foot instances out of the COCO dataset is labeled using the Clickworker platform. It is split up with 14K annotations from the COCO training set and 545 from the validation set. A total of 6 foot keypoints are labeled. We consider the 3D coordinate of the foot keypoints rather than the surface position. For instance, for the exact toe positions, we label the area between the connection of the nail and skin, and also take depth into consideration by labeling the center of the toe rather than the surface.

现有人类姿态数据集的身体部位标注类型较为有限。MPII数据集仅标注了脚踝、膝盖、髋部、肩部、肘部、腕部、颈部、躯干及头顶关键点，COCO数据集则额外覆盖了部分面部关键点。上述两类数据集的足部标注均仅局限于脚踝位置。然而，诸如头像重定向（avatar retargeting）、三维人体形状重建（3D human shape reconstruction）等图形学应用，需要大脚趾、脚跟等足部关键点信息。若缺失足部相关标注，此类方法会出现糖纸效应（candy wrapper effect）、地面穿透（floor penetration）以及足部滑移（foot skate）等问题。为解决上述问题，研究人员通过Clickworker平台对COCO数据集中的部分足部实例进行了标注。该标注子集涵盖COCO训练集中的1.4万条标注与验证集中的545条标注，共计标注6个足部关键点。本次标注采用足部关键点的三维坐标而非表面位置。例如，针对精确的脚趾位置，我们标注指甲与皮肤连接处的区域，并通过标注脚趾中心而非表面位置来考量深度信息。

搜集汇总

背景与挑战

背景概述

Human Foot Keypoint数据集是一个专注于足部关键点标注的2D关键点数据集，包含从COCO数据集中标注的14K训练集和545验证集样本，标注了6个足部关键点（如大脚趾和脚跟），用于解决3D人体形状重建中的足部定位问题。