跳台滑雪运动员姿态估计及运动分析代码数据集

本数据集主要面向跳台滑雪运动员姿态估计及运动分析研究、冬奥运动会跳台滑雪远动员运动过程识别需求建设，主要研究了自下而上的人体姿态估计方法在跳台滑雪运动中的应用，给出了一种基于门控通道变换（Gated Channel Transformation，GCT）的人体姿态估计方法，研究了自上而下的人体姿态估计方法在跳台滑雪运动中的应用，给出了一种基于跨通道信息交互的人体姿态估计方法。通过像素级融合、惯性传感数据归一化处理、二维卷积神经网络和特征提取及融合等技术进行跳台滑雪运动识别。所有数据均采集自安徽大学电子信息工程学院，使用Python编程语言来进行跳台滑雪运动阶段识别实验。

This dataset is developed for research on pose estimation and motion analysis of ski jumping athletes, as well as the demand for motion process recognition of ski jumping athletes in the Winter Olympic Games. It mainly explores the application of bottom-up human pose estimation methods in ski jumping, and proposes a human pose estimation method based on Gated Channel Transformation (GCT). Additionally, it investigates the application of top-down human pose estimation methods in ski jumping, and presents a human pose estimation method based on cross-channel information interaction. Ski jumping motion recognition is implemented via multiple technologies including pixel-level fusion, inertial sensing data normalization, 2D convolutional neural networks, feature extraction and fusion. All data are collected from the School of Electronic and Information Engineering, Anhui University, and Python programming language is utilized for ski jumping motion phase recognition experiments.

搜集汇总

背景与挑战

背景概述

该数据集专注于跳台滑雪运动员的姿态估计及运动分析研究，采用自下而上和自上而下的人体姿态估计方法，结合像素级融合和惯性传感数据归一化处理等技术。数据来源于安徽大学电子信息工程学院，实验使用Python编程语言进行。