Research On Fusing Pose Estimation For Still Image Action Recongnition

The research of action recognition has been become popular increasingly in the field of artificial intelligence,and receive more and more attention.Action recognition include action recognition of still images and video action recognition.In this paper,action recognition for static images were studied and analyzed in detail.Still image has many potential applications in the fields of image search,personal photo album management and human-computer interaction.In this paper,an action recognition method is proposed,which fuses the multiple pose estimation features under the multiple action model.These feature will be used to match with the template.This method can adapt to occlusion or other complex situations.To adapt to occlusion or other complex situations,an action recognition method is proposed which fuses multiple pose estimation features.First,Multiple pose features will be obtained using multiple action models.Each pose feature information includes key point positions and pose scores.Then distinguishing key points are extracted from all train images and computing relative distances between point pairs.An action template is built using all features of the train images of the action.Finally multiple feature information consistent with multiple action templates are extracted from each test image from multiple pose features.Multiple features of the test image are matched with the corresponding action template and then matched values are optimized using pose scores.The experimental results have shown that the average accuracy of the proposed method is approximately 2% better than some other state-of-the-art methods on VOC 2011-val set,and is approximately 6% better than some other state-of-the-art methods on Stanford 40 actions set.By fusing multiple pose features,the proposed method can adapt to occlusion and other complex situations and improve average recognition accuracy.