A 46.1 fps Global Matching Optical Flow Estimation Processor for Action Recognition in Mobile Devices.
Juhyoung LeeChanghyeon KimSungpill ChoiDongjoo ShinSanghoon KangHoi-Jun YooPublished in: ISCAS (2018)
Keyphrases
- action recognition
- optical flow estimation
- optical flow
- feature matching
- human actions
- activity recognition
- flow field
- action classification
- computer vision
- motion field
- matching algorithm
- motion model
- stereo matching
- multiscale
- image matching
- coarse to fine
- motion estimation
- frame rate
- velocity field
- graph matching
- recognizing human actions
- vector field
- multi frame
- human activities
- motion segmentation
- template matching
- keypoints
- image quality
- image sequences