Dense object tracking, the ability to localize specific object points with pixel-level accuracy, is an important computer vision task with numerous downstream applications in robotics. Existing approaches either compute dense keypoint embeddings in a single forward pass, meaning the model is trained to track everything at once, or allocate their full capacity to a sparse predefined set of points, trading generality for accuracy. In this paper we explore a middle ground based on the observation that the number of relevant points at a given time are typically relatively few, e.g. grasp points on a target object. Our main contribution is a novel architecture, inspired by few-shot task adaptation, which allows a sparse-style network to condition on a keypoint embedding that indicates which point to track. Our central finding is that this approach provides the generality of dense-embedding models, while offering accuracy significantly closer to sparse-keypoint approaches. We present results illustrating this capacity vs. accuracy trade-off, and demonstrate the ability to zero-shot transfer to new object instances (within-class) using a real-robot pick-and-place task.
Few-Shot Keypoint Detection as Task Adaptation via Latent Embeddings
2022-07-12
In: Proceedings - IEEE International Conference on Robotics and Automation. (pp. pp. 1251-1257). IEEE: Philadelphia, PA, USA. (2022)
Paper
Elektronische Ressource
Englisch
DDC: | 629 |
MONOCULAR 2D SEMANTIC KEYPOINT DETECTION AND TRACKING
Europäisches Patentamt | 2023
|Performance Evaluation of 3D Keypoint Detectors
British Library Online Contents | 2013
|Representation Learning for Robot Keypoint Detection using Prior Kinematic Knowledge
Deutsches Zentrum für Luft- und Raumfahrt (DLR) | 2022
|Keypoint Detection and Local Feature Matching for Textured 3D Face Recognition
British Library Online Contents | 2008
|