Stream implementation of the flux tensor motion flow algorithm using GStreamer and CUDA.
Dardo D. KleinerKannappan PalaniappanGunasekaran S. SeetharamanPublished in: AIPR (2016)
Keyphrases
- parallel implementation
- times faster
- computational complexity
- preprocessing
- k means
- learning algorithm
- hardware implementation
- segmentation algorithm
- simulated annealing
- optimal solution
- probabilistic model
- real time
- image sequences
- detection algorithm
- space time
- path planning
- configuration space
- np hard
- global motion
- motion segmentation
- motion analysis
- high order
- search space
- optical flow
- higher order
- dynamic programming