Description
We have discussed efficient OMP and K-SVD implementations which vastly improve over the straightforward ones. The Batch-OMP method is useful for a wide range of sparsity-based techniques which require large sets of signals to be coded. The Approximate K-SVD implementation, which employs Batch-OMP as its sparse-coding method, also improves on the dictionary update step by using a significantly faster and more memory efficient atom update computation.
Basis rotation algorithm for approximating the maximum likelihood filters in natural images
Reviews
There are no reviews yet.