A low-complexity activity classification algorithm with optimized selection of accelerometric features.
Matteo GiubertiGianluigi FerrariPublished in: J. Ambient Intell. Smart Environ. (2016)
Keyphrases
- low complexity
- classification algorithm
- classification method
- input features
- class labels
- motion estimation
- training phase
- k nearest neighbor
- computational complexity
- support vector machine
- naive bayes
- accurate classification
- feature extraction
- distributed video coding
- knn
- feature set
- learning algorithm
- classification accuracy
- feature space
- multiple description coding
- lower complexity
- real time
- resource allocation
- prior knowledge
- feature vectors
- neural network