A Hierarchical Segmentation Approach with Convolution-Recursive Deep Learning for 3D Multi-Object Recognition under Partial Occlusion Conditions.
Somar BoubouTatsuo NarikiyoMichihiro KawanishiPublished in: MVA (2019)
Keyphrases
- deep learning
- partial occlusion
- hierarchical segmentation
- object class
- object recognition
- unsupervised learning
- coarse to fine
- image database
- object detection
- mean shift
- machine learning
- object tracking
- visual tracking
- lighting conditions
- computer vision
- mental models
- image processing
- d objects
- image features
- feature extraction
- image sequences
- ground truth
- training data