Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Detection of Autonomous Shuttles in Urban Traffic Images Using Adaptive Residual Context

Mohamed Aziz Younes, Nicolas Saunier, Guillaume-Alexandre Bilodeau

Abstract

The progressive automation of transport promises to enhance safety and sustainability through shared mobility. Like other vehicles and road users, and even more so for such a new technology, it requires monitoring to understand how it interacts in traffic and to evaluate its safety. This can be done with fixed cameras and video object detection. However, the addition of new detection targets generally requires a fine-tuning approach for regular detection methods. Unfortunately, this implementation strategy will lead to a phenomenon known as catastrophic forgetting, which causes a degradation in scene understanding. In road safety applications, preserving contextual scene knowledge is of the utmost importance for protecting road users. We introduce the Adaptive Residual Context (ARC) architecture to address this. ARC links a frozen context branch and trainable task-specific branches through a Context-Guided Bridge, utilizing attention to transfer spatial features while preserving pre-trained representations. Experiments on a custom dataset show that ARC matches fine-tuned baselines while significantly improving knowledge retention, offering a data-efficient solution to add new vehicle categories for complex urban environments.

Detection of Autonomous Shuttles in Urban Traffic Images Using Adaptive Residual Context

Abstract

The progressive automation of transport promises to enhance safety and sustainability through shared mobility. Like other vehicles and road users, and even more so for such a new technology, it requires monitoring to understand how it interacts in traffic and to evaluate its safety. This can be done with fixed cameras and video object detection. However, the addition of new detection targets generally requires a fine-tuning approach for regular detection methods. Unfortunately, this implementation strategy will lead to a phenomenon known as catastrophic forgetting, which causes a degradation in scene understanding. In road safety applications, preserving contextual scene knowledge is of the utmost importance for protecting road users. We introduce the Adaptive Residual Context (ARC) architecture to address this. ARC links a frozen context branch and trainable task-specific branches through a Context-Guided Bridge, utilizing attention to transfer spatial features while preserving pre-trained representations. Experiments on a custom dataset show that ARC matches fine-tuned baselines while significantly improving knowledge retention, offering a data-efficient solution to add new vehicle categories for complex urban environments.
Paper Structure (14 sections, 5 equations, 6 figures, 1 table)

This paper contains 14 sections, 5 equations, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Object Detection in Shared Autonomous Mobility
  4. Continual Learning and Catastrophic Forgetting
  5. Attention-Guided Feature Fusion
  6. Methodology
  7. Channel Attention
  8. Spatial Gating
  9. Residual Project into the Task-specific Branch
  10. Experiments
  11. Dataset and Evaluation Metrics
  12. Implementation Details and Model Training
  13. Comparison with Baseline Solutions
  14. Conclusion

Figures (6)

  • Figure 1: An example of the ARC model outputting joint detections of the original dataset classes (in red) and a new task specific class (in green).
  • Figure 2: The ARC architecture: A frozen YOLO11 branch transfers features to trainable task-specific branches via a Context-Guided Bridge. The bridge employs sequential Channel Attention and Spatial Gating, fused via a learnable residual connection to enhance target detection. Several task specific branches can be added to handle new classes
  • Figure 3: Inside the Context-Guided Bridge. This bridge utilizes Channel Attention and Spatial Gating to refine features before injecting them into the task-specific stream through a learnable residual connection Alpha ($\alpha$).
  • Figure 4: Annotated samples from our custom dataset, showcasing diverse environments; points of views, and lighting conditions.
  • Figure 5: Statistical analysis of the dataset. Top: Instance counts, normalized areas (scale variability), and aspect ratios. Bottom: Spatial distribution of box centers (spatial bias) and width/height histograms.
  • ...and 1 more figures