Summary of Run-time Monitoring Of 3d Object Detection in Automated Driving Systems Using Early Layer Neural Activation Patterns, by Hakan Yekta Yatbaz et al.
Run-time Monitoring of 3D Object Detection in Automated Driving Systems Using Early Layer Neural Activation Patterns
by Hakan Yekta Yatbaz, Mehrdad Dianati, Konstantinos Koufos, Roger Woodman
First submitted to arxiv on: 11 Apr 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Robotics (cs.RO)
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 Monitoring object detection errors in automated driving systems (ADS) is crucial for ensuring safety. While DNN-based detectors have improved, they remain susceptible to errors, particularly in 3D object detection. Traditional integrity monitoring methods rely on activation patterns from the final layer of the detector’s backbone, which may not suffice for complex 3D data. This paper investigates using activation patterns from various layers to introspect 3D object detectors, analyzing Kitti and NuScenes datasets with PointPillars and CenterPoint detectors. Results show that earlier layers’ patterns enhance error detection performance but increase computational complexity. To address real-time operation requirements in ADS, this study introduces a novel introspection method combining multiple layer activation patterns. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making sure object detection systems for self-driving cars work correctly. These systems use special computer programs to detect objects like cars or pedestrians on the road. While these systems have gotten better, they can still make mistakes, especially when it comes to detecting things in 3D space (like the distance between objects). The authors of this paper looked at how to improve error detection for these systems and found that using information from earlier stages of the program can help detect errors better. However, this method requires more computer power. To fix this problem, they came up with a new way to detect errors that works faster. |
Keywords
» Artificial intelligence » Object detection
