Fast human-animal detection from highly cluttered camera-trap images using joint background modeling and deep learning classification.
Hayder YousifJianhe YuanRoland KaysZhihai HePublished in: ISCAS (2017)
Keyphrases
- background modeling
- deep learning
- complex background
- image classification
- multiple images
- varying illumination
- input image
- image features
- unsupervised learning
- feature extraction
- object detection
- three dimensional
- lighting conditions
- object recognition
- machine learning
- pattern recognition
- test images
- feature space
- background subtraction
- foreground objects
- single image
- graph cuts
- moving camera
- light source
- semi supervised
- feature vectors
- training set
- computer vision