Drivable Dirt Road Region Identification Using Image and Point Cloud Semantic Segmentation Fusion.
Jihun ParkHyunsung YooYooseung WangPublished in: IEEE Trans. Intell. Transp. Syst. (2022)
Keyphrases
- semantic segmentation
- point cloud
- input image
- superpixels
- region of interest
- multiscale
- conditional random fields
- image regions
- image classification
- scene classification
- single image
- image retrieval
- image representation
- test images
- image features
- image data
- bounding box
- feature points
- weakly supervised
- structure from motion
- image content
- image structure
- target object
- pixel level
- segmentation algorithm
- image segmentation
- sparse reconstruction
- image matching
- image registration
- markov random field
- super resolution