Geometric Feature
Geometric features, encompassing shape, spatial relationships, and other structural properties, are crucial for various tasks across computer vision, machine learning, and other fields. Current research focuses on leveraging these features through advanced architectures like graph neural networks and transformers, often combined with visual features or other modalities, to improve the accuracy and robustness of applications such as object detection, pose estimation, and trajectory prediction. This work is significant because effective utilization of geometric features enhances the understanding and analysis of complex data, leading to improvements in diverse applications ranging from robotics and autonomous driving to medical image analysis and engineering design.
Papers
GeoBiked: A Dataset with Geometric Features and Automated Labeling Techniques to Enable Deep Generative Models in Engineering Design
Phillip Mueller, Sebastian Mueller, Lars Mikelsons
Spotlight Text Detector: Spotlight on Candidate Regions Like a Camera
Xu Han, Junyu Gao, Chuang Yang, Yuan Yuan, Qi Wang