Speeding-up and compression convolutional neural networks by low-rank decomposition without fine-tuning.
Meng ZhangFei LiuDongpeng WengPublished in: J. Real Time Image Process. (2023)
Keyphrases
- fine tuning
- low rank
- convolutional neural networks
- tensor decomposition
- linear combination
- convex optimization
- matrix factorization
- missing data
- matrix completion
- high order
- singular value decomposition
- rank minimization
- high dimensional data
- matrix decomposition
- fine tuned
- low rank matrix
- image compression
- semi supervised
- trace norm
- minimization problems
- robust principal component analysis
- singular values
- small number
- incomplete data
- missing values
- collaborative filtering
- image sequences
- binary matrix
- non rigid structure from motion
- image processing
- data sets