Deep MANTA: A Coarse-to-Fine Many-Task Network for Joint 2D and 3D Vehicle Analysis from Monocular Image.
Florian ChabotMohamed ChaouchJaonary RabarisoaCéline TeulièreThierry ChateauPublished in: CVPR (2017)
Keyphrases
- coarse to fine
- multiscale
- hierarchical segmentation
- image pyramid
- image analysis
- matching scheme
- scan line
- image data
- multiresolution
- hierarchical representation
- input image
- object detection
- object recognition
- image segmentation
- template matching
- pose estimation
- image registration
- image matching
- edge detection
- unsupervised image segmentation
- segmentation method
- face detection
- pixel level
- image retrieval
- feature correspondences
- image database
- feature space
- kernel estimation
- dynamic programming