RNN BCJR: a fully trainable version of the additive BCJR algorithm.
Guido MontorsiBarbara RipaniPublished in: ICC (2023)
Keyphrases
- learning algorithm
- dynamic programming
- experimental evaluation
- optimization algorithm
- preprocessing
- significant improvement
- computational complexity
- detection algorithm
- cost function
- video sequences
- data sets
- nearest neighbor
- worst case
- high accuracy
- memory requirements
- monte carlo
- single pass
- recurrent neural networks
- segmentation algorithm
- theoretical analysis
- simulated annealing
- computational cost
- objective function
- computationally efficient
- np hard
- k means
- search space
- search algorithm
- selection algorithm
- image processing