Search

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. [ABSTRACT FROM AUTHOR]

Steady-state visual evoked potential (SSVEP) is one of the typical stimulation paradigms of brain-computer interface (BCI). It has become a research approach to improve the performance of human-computer interaction, because of its advantages including multiple objectives, less recording electrodes for electroencephalogram (EEG) signals, and strong anti-interference capacity. Traditional SSVEP using light flicker stimulation may cause visual fatigue with a consequent reduction of recognition accuracy. To avoid the negative impacts on the brain response caused by prolonged strong visual stimulation for SSVEP, steady-state motion visual evoked potential (SSMVEP) stimulation method was used in this study by an equal-luminance colored ring-shaped checkerboard paradigm. The movement patterns of the checkerboard included contraction and expansion, which produced less discomfort to subjects. Feature recognition algorithms based on power spectrum density (PSD) peak was used to identify the peak frequency on PSD in response to visual stimuli. Results demonstrated that the equal-luminance red-green stimulating paradigm within the low frequency spectrum (lower than 15 Hz) produced higher power of SSMVEP and recognition accuracy than black-white stimulating paradigm. PSD-based SSMVEP recognition accuracy was 88.15±6.56%. There was no statistical difference between canonical correlation analysis (CCA) (86.57±5.37%) and PSD on recognition accuracy. This study demonstrated that equal-luminance colored ring-shaped checkerboard visual stimulation evoked SSMVEP with better SNR on low frequency spectrum of power density and improved the interactive performance of BCI. [ABSTRACT FROM AUTHOR]