Integration of top-down and bottom-up visual processing using a recurrent convolutional-deconvolutional neural network for semantic segmentation.
Byung Wan KimYoungbin ParkIl Hong SuhPublished in: Intell. Serv. Robotics (2020)
Keyphrases
- semantic segmentation
- visual processing
- neural network
- feed forward
- visual cortex
- recurrent neural networks
- conditional random fields
- computational model
- superpixels
- visual information
- object categories
- weakly supervised
- artificial neural networks
- visual attention
- object recognition
- back propagation
- object classes
- hidden layer
- natural images
- hidden markov models