AIFB 06-2022 Bachelor or Master Thesis: Specialized Evaluation Metrics for Perception Tasks in Autonomous Driving
Organisationseinheit
Institut für Angewandte Informatik und Formale Beschreibungsverfahren (AIFB)
Ihre Aufgaben
In computer vision, we often deal with metrics that evaluate results irrelevant of the task at hand. However, for the field of autonomous driving, this method is flawed - you don’t necessarily care about misclassifications etc. if they do not affect the current trajectory. In this thesis, based on a 3D voxel space, you will develop a relevance metric that is adapted to the task of autonomous driving.
The Topic:
- You will perform state-of-the-art research on evaluation metrics for the task of autonomous driving
- You will implement a vanilla version of a path planner for the autonomous vehicle in the CARLA simulation environment
- You will represent the world in a 3D voxel space
- You will develop a method that assigns a weight to each voxel, enabling any perception task (object detection, anomaly detection, …) to receive a weighted version of their output, adapted to the driving task
- You will create appealing image and video visualizations
Eintrittstermin
as soon as possible
Ihre Qualifikation
Your Skills:
- You study Computer Science or a related course of study
- You are deeply interested in topics such as Autonomous Driving, Robotics, Deep Learning or Computer Vision
- You are able to read and write scientific texts in English
- You are fluid in Python and familiar with Linux
- You show an above-average degree of initiative and commitment as well as a thorough way of working
Das bieten wir Ihnen
- You get exciting insights into our research and gain valuable practical experience
- We use the latest hardware and software. Together with us you work in first-class laboratories (on-site or remotely)
- Regular and extensive support: Weekly 1:1 meetings, bi-weekly student group meetings, monthly 1:1 strategy meetings
- Collaboration with other students to get tips, learn together, and fix issues quickly High-quality theses will be published on KITopen, with the code on GitHub
- We aim to publish this work in an IEEE paper with shared first authorship
Contract duration
6 Months
Contact person in line-management
For further information, please contact Daniel Bogdoll, email: daniel.bogdoll@kit.edu.
Application
Shoot me an e-mail at daniel.bogdoll@kit.edu with your CV, grades, and a few sentences why you are interested. No cover letter necessary.
Recognized severely disabled persons will be preferred if they are equally qualified.