Fully Automated Annotation With Noise-Masked Visual Markers for Deep-Learning-Based Object Detection.
Takuya KiyokawaKeita TomochikaJun TakamatsuTsukasa OgasawaraPublished in: IEEE Robotics Autom. Lett. (2019)
Keyphrases
- fully automated
- deep learning
- object detection
- labor intensive
- fully automatic
- unsupervised learning
- unsupervised feature learning
- weakly supervised
- object class
- machine learning
- mental models
- multi class
- computer vision
- visual features
- image annotation
- object segmentation
- object categories
- active learning
- image retrieval
- image classification
- semi automatic