Abstract
In the process of human behavior recognition, the traditional dense optical flow method has too many pixels and too much overhead, which limits the running speed. This paper proposed a method combing YOLOv3 (You Only Look Once v3) and local optical flow method. Based on the dense optical flow method, the optical flow modulus of the area where the human target is detected is calculated to reduce the amount of computation and save the cost in terms of time. And then, a threshold value is set to complete the human behavior identification. Through design algorithm, experimental verification and other steps, the walking, running and falling state of human body in real life indoor sports video was identified. Experimental results show that this algorithm is more advantageous for jogging behavior recognition.
Highlights
As a sub-field of artificial intelligence technology, computer vision and deep learning have developed rapidly in recent years
This paper proposed a method combing YOLOv3 (You Only Look Once v3) and local optical flow method
The local optical flow method based on YOLOv3 is to localize the dense optical flow method on the basis of YOLOv3 algorithm to realize the purpose of saving running time and speeding up running speed
Summary
As a sub-field of artificial intelligence technology, computer vision and deep learning have developed rapidly in recent years. Human behavior recognition technology based on video sequence has developed from the earliest traditional classification method based on manual design features to the current method based on deep learning automatic feature extraction. The former requires manual design of features and classification by classifier. The sparse optical flow method in the traditional optical flow method only calculates individual points or points with special significance in image recognition, while the dense optical flow method proposed by Gunnar Farneback [5] calculates the dense optical flow field of the entire image by calculating the motion translation model of each pixel on the image This method can obtain the feature of video image at pixel level through dense optical flow field, so the effect of human motion recognition based on it is obviously better than that based on sparse optical flow. Wang et al introduced a new continuous optical flow framework to capture pixel dynamics by representing a group of continuous RGB frames sequentially through
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call