Lane Marking Detection Algorithms Evaluation for Safe Autonomous Vehicles

Principle Goal: Summarize and supplement state-of-the-art evaluation metrics for lane marking detection algorithms.
Are you fascinated by autonomous vehicles and eager to play a pivotal role in their development? We're seeking enthusiastic individuals who are interested in contributing to one of the key aspects of autonomous driving: lane marking detection.
Lane marking detection is crucial for the lane-keeping and lane-changing capabilities of autonomous vehicles (AVs). Miss detection or inaccurate detection could lead to severe traffic incidents. Therefore, it is essential to evaluate lane marking detection algorithms thoroughly using appropriate metrics. Metrics like the false negative rate can be computed offline to gauge algorithm performance, while others are applicable in real-time during AV operation. Despite the importance of these metrics, a systematic summarization and comparison are missing. Moreover, there are limited metrics dedicated to real-time safety monitoring of lane marking detection. This project aims to bridge this gap by investigating state-of-the-art evaluation metrics and eventually proposing new metrics for real-time safety monitoring of lane marking detection algorithms.
To accomplish this objective, the project will require the following:
1) Familiarize yourself with current lane marking detection algorithms and select one as your study object.
2) Summarize and classify state-of-the-art evaluation metrics for lane marking detection algorithms.
3) Identify the limitations of current existing metrics.
4) Design and implement new evaluation metrics for real-time safety monitoring of lane marking detection algorithms.
5) Evaluate the metrics and summarize your results.
To complete the tasks, the following knowledge/skills could be useful:
1) Have Python coding experience.
2) Have expertise in deep learning and image processing.
3) A strong understanding of autonomous driving and lane marker detection is preferred.
4) Having safety engineering knowledge is preferred.

Note: this topic will be supervised by our new academic, Dr. Cheng Wang once he starts working at the EPS.