Identifying and Extracting Football Features from Real-World Media Sources using Only Synthetic Training Data.
Jose Cerqueira FernandesBenjamin KenwrightPublished in: CoRR (2022)
Keyphrases
- real world
- training data
- classification models
- data sets
- automatically extracted
- image features
- low level
- classification accuracy
- extracting features
- prior knowledge
- co occurrence
- multimedia
- databases
- machine learning
- feature extraction
- case study
- data sources
- feature selection
- wide range
- supervised learning
- information sources
- object detection
- feature values
- data mining
- test data
- false positives
- feature set
- learning algorithm
- domain knowledge
- feature vectors
- training set
- feature space
- object recognition