Abstract:Human action recognition using 3D pose data has gained a growing interest in the field of computer robotic interfaces and pattern recognition since the availability of hardware to capture human pose. In this paper, we propose a fast, simple, and powerful method of human action recognition based on human kinematic similarity. The key to this method is that the action descriptor consists of joints position, angular velocity and angular acceleration, which can meet the different individual sizes and eliminate the complex normalization. The angular parameters of joints within a short sliding time window (approximately 5 frames) around the current frame are used to express each pose frame of human action sequence. Moreover, three modified KNN (k-nearest-neighbors algorithm) classifiers are employed in our method: one for achieving the confidence of every frame in the training step, one for estimating the frame label of each descriptor, and one for classifying actions. Additional estimating of the frame's time label makes it possible to address single input frames. This approach can be used on difficult, unsegmented sequences. The proposed method is efficient and can be run in real time. The research shows that many public datasets are irregularly segmented, and a simple method is provided to regularize the datasets. The approach is tested on some challenging datasets such as MSR-Action3D, MSRDailyActivity3D, and UTD-MHAD. The results indicate our method achieves a higher accuracy.

Discriminative Pose Analysis for Human Action Recognition.

Discriminative Hierarchical Part-Based Models for Human Parsing and Action Recognition.

Modelling Human Body Pose for Action Recognition Using Deep Neural Networks

Human Action Recognition with Contextual Constraints Using a RGB-D Sensor

Kpose: A New Representation For Action Recognition

Exploring discriminative pose sub-patterns for effective action classification.

Human Action Recognition Based on Kinematic Similarity in Real Time.

Human Activity Recognition Based on Pose Points Selection

Learning Discriminative Activated Simplices for Action Recognition

Discriminative Estimation of 3D Human Pose Using Gaussian Processes.

B2C-AFM: Bi-Directional Co-Temporal and Cross-Spatial Attention Fusion Model for Human Action Recognition.

An Approach to Pose-Based Action Recognition

Typing Video frames after person detection Pose Tube 2 D Deconv Score fusion RGB action recognition Pose action recognition Pose estimation

ActionPose: Pretraining 3D Human Pose Estimation with the Dark Knowledge of Action

Joint Dynamic Pose Image and Space Time Reversal for Human Action Recognition from Videos

Human Action Recognition Based on Semi-Supervised Discriminant Analysis with Global Constraint

Dual-view 3D Human Pose Estimation Without Camera Parameters for Action Recognition

Oriented Gradients for Human Action Recognition.

Human Action Recognition Using Sparse Representation

Learning a Similarity Metric Discriminatively for Pose Exemplar Based Action Recognition

Pose Adaptive Motion Feature Pooling for Human Action Analysis