Summary of Feature Importance in Pedestrian Intention Prediction: a Context-aware Review, by Mohsen Azarmi et al.
Feature Importance in Pedestrian Intention Prediction: A Context-Aware Review
by Mohsen Azarmi, Mahdi Rezaei, He Wang, Ali Arabian
First submitted to arxiv on: 11 Sep 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Robotics (cs.RO); Image and Video Processing (eess.IV)
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 tackles the challenge of understanding how Deep Neural Networks (DNNs) predict pedestrian crossing intentions for Autonomous Vehicles using Computer Vision. The DNN’s black-box nature makes it difficult to comprehend how input features contribute to final predictions, which hinders informed decisions on feature selection and optimisation. To address this, the authors introduce Context-aware Permutation Feature Importance (CAPFI), a novel approach that leverages subdivided scenario contexts to reduce variance and prevent biased estimations in importance scores. The paper evaluates five neural network architectures for intention prediction across 16 comparable context sets using CAPFI, revealing nuanced differences among models. The findings highlight the critical role of pedestrian bounding boxes and ego-vehicle speed in predicting pedestrian intentions and potential prediction biases due to the speed feature. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research helps autonomous vehicles better understand pedestrians’ intentions. By solving a mystery about how deep learning works, scientists can make better decisions about what features are important for predictions. They developed a new method called CAPFI that looks at different scenarios to figure out which features matter most. The study tested five different neural networks and found that some work better than others in certain situations. This could help create more accurate and reliable prediction models. |
Keywords
» Artificial intelligence » Deep learning » Feature selection » Neural network
