Unsupervised open world recognition in computer vision / eingereicht von Svenja Uhlemeyer, M.Sc. Wuppertal, 29.03.2023
Inhalt
- Acknowledgments
- Foreword
- Contents
- Introduction
- Foundations
- Statistical Learning Theory for Supervised Learning
- Deep Neural Networks
- Deep Learning for Computer Vision
- Learning from Unlabeled Data
- Road Anomalies
- Towards Unsupervised Open World Semantic Segmentation
- Introduction
- Related Works
- Discovery of Unknown Semantic Classes
- Meta Regressor
- Uncertainty Metrics and Prediction Quality Estimation
- Embedding and Clustering of Image Patches
- Novelty Segmentation
- Extension of the Model's Semantic Space
- Experiments
- Conclusion
- Detecting Novelties with Empty Classes
- Introduction
- Related Works
- Method Description
- Adjustments for Semantic Segmentation
- Numerical Experiments
- Conclusion & Outlook
- Conclusion & Outlook
- List of Figures
- List of Tables
- Bibliography