Isometric-Multiclass Dataset (IMCD)
收藏资源简介:
Isometric-Multiclass Dataset (IMCD) 是一个包含超过 3,600 个视频片段的多类别等长运动视频数据集,涵盖六种姿势的正确和错误变化。该数据集由印度理工学院坎普尔分校的研究人员创建,旨在解决家庭锻炼中缺乏实时反馈的问题。IMCD 数据集包含十位不同体型和表演风格的参与者,每个练习都包含了超过 500 次正确和错误类别的数据。数据在室内和室外设置中录制,参与者穿着标准的家居服装,以确保数据集尽可能接近真实的生活场景。IMCD 数据集可用于训练和评估自动反馈系统,以提高家庭锻炼计划的质量和可持续性。
Isometric-Multiclass Dataset (IMCD) is a multi-class isometric motion video dataset comprising over 3,600 video clips, covering correct and incorrect variations of six postures. This dataset was created by researchers from the Indian Institute of Technology Kanpur to address the lack of real-time feedback during home workouts. The IMCD dataset includes ten participants with different body types and performance styles, with each exercise category containing over 500 samples of both correct and incorrect instances. The data was recorded in both indoor and outdoor settings, with participants wearing standard household clothing to ensure the dataset closely approximates real-life scenarios. The IMCD dataset can be utilized to train and evaluate automated feedback systems to improve the quality and sustainability of home exercise programs.

- 1通过印度理工学院坎普尔分校 · 2025年



