Waveplot-based Dataset for Multi-class Human Action Analysis

This dataset comprises an assortment of waveplot images representing diverse human actions. Waveplot images are time-amplitude representations of audio signals that encapsulate the variation of audio amplitude over time. In this dataset, the audio signals correspond to disparate human actions, such as walking, running, jumping, and dancing. The waveplot images are created by plotting the amplitude of the audio signals against time, with each image representing a segment of the audio signal. The dataset is explicitly designed for tasks like human action recognition, classification, segmentation, and detection based on auditory cues. It serves as a valuable resource for training and evaluating machine learning models that analyze human actions predicated on audio signals. The dataset caters well to researchers and practitioners in the disciplines of signal processing, computer vision, and machine learning, who are keen on devising algorithms for human action analysis using audio signals. Crucially, the dataset is annotated with labels that denote the type of human action represented in each waveplot image. This ensures a supervised learning environment conducive for the development and testing of prediction models.本数据集包含一系列表征各类人类动作的波形图（waveplot）图像。波形图是音频信号的时幅表征形式，可完整呈现音频振幅随时间的变化规律。本数据集内的音频信号对应各类不同的人类动作，例如行走、奔跑、跳跃与舞蹈。此类波形图通过将音频信号的振幅随时间轴绘制生成，每张图像对应一段音频信号片段。本数据集专为基于听觉线索的人类动作识别、分类、分割与检测等任务设计，可作为训练与评估基于音频信号分析人类动作的机器学习模型的宝贵资源。本数据集能够很好地适配信号处理、计算机视觉与机器学习领域中，致力于利用音频信号开发人类动作分析算法的研究人员与从业者的需求。至关重要的是，本数据集已为每张波形图标注了其所表征的人类动作类别标签，这为预测模型的开发与测试提供了适宜的监督学习环境。