Summary of A Safety-adapted Loss For Pedestrian Detection in Automated Driving, by Maria Lyssenko et al.
A Safety-Adapted Loss for Pedestrian Detection in Automated Driving
by Maria Lyssenko, Piyush Pimplikar, Maarten Bieshaar, Farzad Nozarian, Rudolph Triebel
First submitted to arxiv on: 5 Feb 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Computer Vision and Pattern Recognition (cs.CV)
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 This paper proposes a novel approach to improve object detection in automated driving (AD) scenarios. The existing methods focus on identifying safety-critical vulnerable road users (VRU) and annotating them with risk scores, but they don’t consider the safety factor during deep neural network (DNN) training. As a result, state-of-the-art DNNs equally penalize all misdetections, including critical failures. To mitigate this issue, the authors suggest a safety-aware training strategy that leverages estimated per-pedestrian criticality scores during training. They utilize reachability set-based time-to-collision (TTC-RSB) metric and distance information to quantify the criticality of pedestrians. Experimental results using RetinaNet and FCOS on nuScenes dataset show that models trained with this approach reduce misdetection of critical pedestrians without compromising performance for general cases. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about making cars safer by improving how they detect people. The current way of doing it doesn’t take into account how important each person is to their safety. So, the authors came up with a new idea to make the car’s brain think more about keeping people safe while still being good at detecting them. They use special information like how close someone is to getting hit and where they are on the road. The test results show that this new way makes cars better at detecting important people without making them worse at detecting regular people. |
Keywords
* Artificial intelligence * Neural network * Object detection
