Comparing Optical Flow and Deep Learning to Enable Computationally Efficient Traffic Event Detection with Space-Filling Curves

TL;DR

This paper compares optical flow and deep learning methods for efficient traffic event detection using space-filling curves, demonstrating their respective strengths in specificity and sensitivity on large-scale datasets.

Contribution

It introduces a novel approach combining optical flow and deep learning with space-filling curves for real-time traffic event detection, validated on large-scale datasets.

Findings

Optical flow approach has higher specificity and fewer false positives.

Deep learning approach offers higher sensitivity for event detection.

Both methods achieve comparable processing speeds suitable for real-time use.

Abstract

Gathering data and identifying events in various traffic situations remains an essential challenge for the systematic evaluation of a perception system's performance. Analyzing large-scale, typically unstructured, multi-modal, time series data obtained from video, radar, and LiDAR is computationally demanding, particularly when meta-information or annotations are missing. We compare Optical Flow (OF) and Deep Learning (DL) to feed computationally efficient event detection via space-filling curves on video data from a forward-facing, in-vehicle camera. Our first approach leverages unexpected disturbances in the OF field from vehicle surroundings; the second approach is a DL model trained on human visual attention to predict a driver's gaze to spot potential event locations. We feed these results to a space-filling curve to reduce dimensionality and achieve computationally efficient event retrieval. We systematically evaluate our concept by obtaining characteristic patterns for both approaches from a large-scale virtual dataset (SMIRK) and applied our findings to the Zenseact Open Dataset (ZOD), a large multi-modal, real-world dataset, collected over two years in 14 different European countries. Our results yield that the OF approach excels in specificity and reduces false positives, while the DL approach demonstrates superior sensitivity. Both approaches offer comparable processing speed, making them suitable for real-time applications.