Teachers in concordance for pseudo-labeling of 3D sequential data
| Autor: | Awet Haileslassie Gebrehiwot, Patrik Vacek, David Hurych, Karel Zimmermann, Patrick Perez, Tomas Svoboda |
|---|---|
| Rok vydání: | 2022 |
| Předmět: |
FOS: Computer and information sciences
Control and Optimization Mechanical Engineering Computer Vision and Pattern Recognition (cs.CV) I.4.6 Biomedical Engineering Computer Science - Computer Vision and Pattern Recognition I.4.8 Computer Science Applications Human-Computer Interaction Computer Science - Robotics 68T07 Artificial Intelligence Control and Systems Engineering Computer Vision and Pattern Recognition Robotics (cs.RO) |
| DOI: | 10.48550/arxiv.2207.06079 |
| Popis: | Automatic pseudo-labeling is a powerful tool to tap into large amounts of sequential unlabeled data. It is specially appealing in safety-critical applications of autonomous driving, where performance requirements are extreme, datasets are large, and manual labeling is very challenging. We propose to leverage sequences of point clouds to boost the pseudolabeling technique in a teacher-student setup via training multiple teachers, each with access to different temporal information. This set of teachers, dubbed Concordance, provides higher quality pseudo-labels for student training than standard methods. The output of multiple teachers is combined via a novel pseudo label confidence-guided criterion. Our experimental evaluation focuses on the 3D point cloud domain and urban driving scenarios. We show the performance of our method applied to 3D semantic segmentation and 3D object detection on three benchmark datasets. Our approach, which uses only 20% manual labels, outperforms some fully supervised methods. A notable performance boost is achieved for classes rarely appearing in training data. Comment: This work has been submitted to the IEEE for publication |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|