基于跨域字典学习算法的人体行为识别

摘要

将一种跨域字典学习算法应用于人体行为识别中，通过引入辅助域数据集，与原始训练集(目标域)共同进行字典学习，获得字典对，进而得到动作类的稀疏编码，有效扩充了训练集的类内多样性。该算法为字典学习与训练分类相结合的学习框架，可利用字典对学习过程中的重建误差进行分类。实验在 MATLAB 仿真条件下进行，将 UCF YouTube 数据集作为原始训练集，将 HMDB51数据集作为辅助域数据集，选取两个数据集动作类别一致的7个动作，根据提出的算法流程进行识别。将该方法与其他两种人体行为识别算法进行对比。结果表明，该方法识别率显著提高，证明了跨域字典学习算法在人体行为识别上的有效性。%A cross-domain dictionary learning algorithm was applied to the recognition of human actions.By introducing the auxiliary domain data set,the algorithm learned from the original training set (target domain)together with the dictionary to obtain the dictionary pair,and then obtain the sparse coding of the action class,and effectively expanded diversity of training set.The algorithm is a combination of learning framework of dictionary learing and training classification,which can be used to classify the reconstruction errors in the learning process.The experiment was carried out under the condition of MATLAB simulation.The UCF YouTube data sets were regarded as the original training set,and the HMDB5 1 data sets were regarded as an auxiliary domain data set.7 coherent human actions of the two data sets were selected to do the work of recognition according to the flow chart of the proposed algorithm.The cross-domain dictionary learning algorithm was compared with another two kinds of human actions recognition algorithms.The results show that the recognition rate of the proposed algorithm is significantly improved.It is proved that the cross-domain dictionary learning algorithm is effective in human actions recognition.