Slow Feature Preprocessing in Deep Neural Networks for Wearable Sensor-Based Locomotion Recognition.
Lukas GünthermannLloyd PellattDaniel RoggenPublished in: PerCom Workshops (2022)
Keyphrases
- preprocessing
- neural network
- preprocessing stage
- feature extraction
- pattern recognition
- preprocessing step
- recognition accuracy
- gesture recognition
- recognition rate
- activity recognition
- object recognition
- image features
- feature vectors
- feature fusion
- recognition process
- sensor networks
- artificial neural networks
- genetic algorithm
- augmented reality
- recognition algorithm
- neural nets
- character recognition
- human computer interaction
- post processing
- improve the recognition accuracy
- automatic recognition
- sensor data
- recurrent neural networks
- degrees of freedom
- neural network model
- self organizing maps
- back propagation
- unsupervised learning
- mobile robot
- feed forward
- human activities
- ambient intelligence
- action recognition
- feature set
- low cost
- preprocessing phase
- feature extractor
- feature space
- extended gaussian image