A general, novel methodology for studying the generation of finite precision error in any algorithm.
Constantin PapaodysseusConstantinos ChalatsisDimitris ArabadjisSolomon ZannosFotios GiannopoulosPanayiotis RousopoulosE. KalfaPublished in: ISSPIT (2013)
Keyphrases
- preprocessing
- detection algorithm
- learning algorithm
- k means
- experimental evaluation
- segmentation algorithm
- high accuracy
- computational complexity
- neural network
- optimization algorithm
- error bounds
- computational cost
- dynamic programming
- special case
- search space
- data sets
- cost function
- prediction error
- error rate
- objective function
- improved algorithm
- times faster
- error function
- matching algorithm
- classification algorithm
- generation algorithm
- theoretical analysis
- expectation maximization
- simulated annealing
- motion estimation
- upper bound
- np hard
- significant improvement
- video sequences
- similarity measure