Dual-mode abnormal skeleton data correction method based on spatio-temporal information and optical flow extraction
The invention relates to a dual-mode abnormal skeleton data correction method based on spatio-temporal information and optical flow extraction, and the method comprises the following steps: S1, obtaining and preprocessing a human body motion video data set, obtaining a training data set, and trainin...
Gespeichert in:
Hauptverfasser: | , , |
---|---|
Format: | Patent |
Sprache: | chi ; eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | The invention relates to a dual-mode abnormal skeleton data correction method based on spatio-temporal information and optical flow extraction, and the method comprises the following steps: S1, obtaining and preprocessing a human body motion video data set, obtaining a training data set, and training a human body posture estimator based on computer vision based on the training data set; s2, generating an initial posture of each video frame for an input video sequence according to the trained human body posture estimator, and performing anomaly detection on the initial postures; s3, respectively searching a latest trusted preorder frame and a latest trusted postorder frame from the front and the back of the abnormal frame, and extracting optical flow information of a continuous sequence between the latest trusted preorder frame and the latest trusted postorder frame; and S4, according to the obtained optical flow information, forward optical flow correction and reverse forward optical flow correction are respe |
---|