SqueezeSeg: Convolutional Neural Nets with Recurrent CRF for Real-Time Road-Object Segmentation from 3D LiDAR Point Cloud.
Bichen WuAlvin WanXiangyu YueKurt KeutzerPublished in: CoRR (2017)
Keyphrases
- point cloud
- neural nets
- object segmentation
- feed forward
- superpixels
- real time
- structure from motion
- object detection
- graph cuts
- image segmentation
- point sets
- neural network
- segmentation algorithm
- segmentation method
- conditional random fields
- laser scanner
- back propagation
- shape prior
- artificial neural networks
- object categories
- point cloud data
- learning tasks
- object classes
- energy function
- vision system
- markov random field
- shape model
- object recognition
- computer vision
- genetic algorithm
- urban areas
- machine learning
- data sets