Incredible progress has been made in human action recognition(HAR),significantly impacting computer vision applications in sports analytics.However,identifying dynamic and complex movements in sports like badminton re...Incredible progress has been made in human action recognition(HAR),significantly impacting computer vision applications in sports analytics.However,identifying dynamic and complex movements in sports like badminton remains challenging due to the need for precise recognition accuracy and better management of complex motion patterns.Deep learning techniques like convolutional neural networks(CNNs),long short-term memory(LSTM),and graph convolutional networks(GCNs)improve recognition in large datasets,while the traditional machine learning methods like SVM(support vector machines),RF(random forest),and LR(logistic regression),combined with handcrafted features and ensemble approaches,perform well but struggle with the complexity of fast-paced sports like badminton.We proposed an ensemble learning model combining support vector machines(SVM),logistic regression(LR),random forest(RF),and adaptive boosting(AdaBoost)for badminton action recognition.The data in this study consist of video recordings of badminton stroke techniques,which have been extracted into spatiotemporal data.The three-dimensional distance between each skeleton point and the right hip represents the spatial features.The temporal features are the results of Fast Dynamic Time Warping(FDTW)calculations applied to 15 frames of each video sequence.The weighted ensemble model employs soft voting classifiers from SVM,LR,RF,and AdaBoost to enhance the accuracy of badminton action recognition.The E2 ensemble model,which combines SVM,LR,and AdaBoost,achieves the highest accuracy of 95.38%.展开更多
为快速准确地计算时间序列数据相似度,引入快速动态时间规划距离(fast dynamic time warping,FDTW),提出了基于FDTW的模糊C均值算法和模糊C中心点聚类算法。FDTW通过对数据序列进行拉伸和压缩匹配时间序列数据,只要形状相同,即使发生时...为快速准确地计算时间序列数据相似度,引入快速动态时间规划距离(fast dynamic time warping,FDTW),提出了基于FDTW的模糊C均值算法和模糊C中心点聚类算法。FDTW通过对数据序列进行拉伸和压缩匹配时间序列数据,只要形状相同,即使发生时间位移也可以准确识别,同时解决了传统DTW计算效率较低的问题。试验结果表明,提出的算法仍能保证聚类的精度。展开更多
基金supported by the Center for Higher Education Funding(BPPT)and the Indonesia Endowment Fund for Education(LPDP),as acknowledged in decree number 02092/J5.2.3/BPI.06/9/2022。
文摘Incredible progress has been made in human action recognition(HAR),significantly impacting computer vision applications in sports analytics.However,identifying dynamic and complex movements in sports like badminton remains challenging due to the need for precise recognition accuracy and better management of complex motion patterns.Deep learning techniques like convolutional neural networks(CNNs),long short-term memory(LSTM),and graph convolutional networks(GCNs)improve recognition in large datasets,while the traditional machine learning methods like SVM(support vector machines),RF(random forest),and LR(logistic regression),combined with handcrafted features and ensemble approaches,perform well but struggle with the complexity of fast-paced sports like badminton.We proposed an ensemble learning model combining support vector machines(SVM),logistic regression(LR),random forest(RF),and adaptive boosting(AdaBoost)for badminton action recognition.The data in this study consist of video recordings of badminton stroke techniques,which have been extracted into spatiotemporal data.The three-dimensional distance between each skeleton point and the right hip represents the spatial features.The temporal features are the results of Fast Dynamic Time Warping(FDTW)calculations applied to 15 frames of each video sequence.The weighted ensemble model employs soft voting classifiers from SVM,LR,RF,and AdaBoost to enhance the accuracy of badminton action recognition.The E2 ensemble model,which combines SVM,LR,and AdaBoost,achieves the highest accuracy of 95.38%.
文摘养殖水体中溶解氧浓度一直是最重要的水质参数之一。为了精准地对水体溶解氧进行调控,提高养殖生产效率,降低养殖风险,该研究考虑外部天气条件对溶解氧的影响以及溶解氧自身的昼夜变化特征,提出一种基于正则化极限学习机(principal component analysis and clustering method optimized regularized extreme learning machine,PC-RELM)的养殖水体溶解氧数据流预测模型。首先,采用主成分分析法判断影响溶解氧浓度的强重要性因子,降低预测模型的数据维度;其次,利用熵权法计算各时刻点的天气环境指数,并利用快速动态时间规整算法(fast dynamic time warping,FastDTW)完成时间序列数据流在不同天气环境下的相似度度量;然后使用k-means算法对时间序列的相似度进行聚类分簇,并基于分簇结果完成正则化极限学习机预测模型的构建,实现溶解氧浓度的估算。最后将PC-RELM模型应用到无锡南泉试验基地养殖池塘的溶解氧预测调控过程中。试验结果表明:PC-RELM的预测均方根误差值(root mean square error,RMSE)为0.9619,与PLS-ELM(partial least squares optimized ELM)、最小二乘支持向量机(least square support vector machine,LSSVM)以及BP神经网络模型进行对比,其RMSE值分别降低了41.54%、54.58%和67.16%。该预测模型可以有效地捕捉不同天气条件下溶解氧的变化特点,具有较高的预测精度和效率。
文摘为快速准确地计算时间序列数据相似度,引入快速动态时间规划距离(fast dynamic time warping,FDTW),提出了基于FDTW的模糊C均值算法和模糊C中心点聚类算法。FDTW通过对数据序列进行拉伸和压缩匹配时间序列数据,只要形状相同,即使发生时间位移也可以准确识别,同时解决了传统DTW计算效率较低的问题。试验结果表明,提出的算法仍能保证聚类的精度。
