Dense outlier detection and open-set recognition based on training with noisy negative images.
Petra BevandicIvan KresoMarin OrsicSinisa SegvicPublished in: CoRR (2021)
Keyphrases
- outlier detection
- object recognition
- three dimensional
- data streams
- detection algorithm
- image retrieval
- feature extraction
- fraud detection
- image features
- density ratio estimation
- credit card fraud detection
- knowledge discovery
- distance based outlier detection
- data mining
- high dimensional datasets
- image classification
- data analysis
- machine learning
- unsupervised learning
- least squares
- density estimation
- training set
- detect outliers
- computer vision
- learning algorithm