RLSTM: A Novel Residual and Recurrent Network for Pedestrian Action Classification.
Soulayma GazzehLiliana Lo PrestiAli DouikMarco La CasciaPublished in: CAIP (2) (2023)
Keyphrases
- action classification
- recurrent networks
- action recognition
- recurrent neural networks
- feed forward
- human actions
- biologically inspired
- bag of words
- pedestrian detection
- neural network
- object detection
- salient features
- static images
- global features
- max margin
- spatio temporal
- computer vision
- back propagation
- text classification
- bag of visual words
- unified model
- image segmentation