Summary of Reliable Student: Addressing Noise in Semi-supervised 3d Object Detection, by Farzad Nozarian et al.
Reliable Student: Addressing Noise in Semi-Supervised 3D Object Detection
by Farzad Nozarian, Shashank Agarwal, Farzaneh Rezaeianaran, Danish Shahzad, Atanas Poibrenski, Christian Müller, Philipp Slusallek
First submitted to arxiv on: 27 Apr 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Semi-supervised 3D object detection can benefit from pseudo-labeling when labeled data is limited. However, recent approaches overlooked the impact of noisy pseudo-labels during training despite efforts to enhance pseudo-label quality through confidence-based filtering. The Reliable Student framework incorporates two complementary approaches to mitigate errors: a class-aware target assignment strategy and reliability weighting strategy. The latter determines weights by querying the teacher network for confidence scores of student-generated proposals. Our approach surpasses the previous state-of-the-art on KITTI 3D object detection benchmark in semi-supervised setting, achieving a 6.2% AP improvement for pedestrian class with only 37 labeled samples available. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Semi-supervised 3D object detection is important because it can help us detect objects in images or videos when we don’t have enough labeled data. Some approaches try to use pseudo-labeling, but they forget that noisy labels can affect the results. The new Reliable Student framework helps solve this problem by using two strategies: one for correct target assignment and another for reducing false positives. This approach works well on a benchmark test called KITTI 3D object detection, and it even beats previous best results. |
Keywords
» Artificial intelligence » Object detection » Semi supervised
