Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Camera anomaly detection and correction for the next generation of mobile robots
The goal of this project is to develop a novel deep-learning algorithm for detecting camera anomalies and restoring the anomaly-free images.
Keywords: Computer Vision, SLAM, Deep learning, Robotics, Autonomous driving
Computer vision is a crucial capability for robotics, where understanding the camera images is necessary to perform numerous tasks, such as localization, navigation, manipulation or surveying of predefined areas. The performance of perception, localization and mapping algorithms is limited, however, because of various anomalies that can occur in everyday scenarios, such as water droplets or dust appearing on the lens or the camera being out of focus. With this project, we would like to address those challenges by detecting such situations and reconstructing the original, anomaly-free image.
The goal of the project is to develop a novel AI-based approach that would permit to detect various kinds of camera anomalies (dust, water, fingerprints, defocus) using a deep neural network and, in a second stage, restore the images using state-of-the-art GANs. Thanks to your efforts, mobile robots will be able to operate even in the toughest conditions. You will have a chance to work with us hand-in-hand on our computer vision and AI algorithms and to deploy your approach on commercial robotic platforms, not just GPUs in the lab. By the end of the project, you will have developed a great amount of experience related to computer vision, neural networks, GANs and embedded GPU/tensor units.
**What We Offer**
- Possibility to contribute to ongoing research in the exciting and quickly developing field of deep computer vision
- Work with a team of enthusiastic roboticists and researchers in a Zurich based robotics startup in collaboration with the Autonomous Systems Lab, one of the largest robotic labs in the world.
- Possibility to deploy your algorithms to different robotic platforms and highly-valued hands-on experience.
Computer vision is a crucial capability for robotics, where understanding the camera images is necessary to perform numerous tasks, such as localization, navigation, manipulation or surveying of predefined areas. The performance of perception, localization and mapping algorithms is limited, however, because of various anomalies that can occur in everyday scenarios, such as water droplets or dust appearing on the lens or the camera being out of focus. With this project, we would like to address those challenges by detecting such situations and reconstructing the original, anomaly-free image.
The goal of the project is to develop a novel AI-based approach that would permit to detect various kinds of camera anomalies (dust, water, fingerprints, defocus) using a deep neural network and, in a second stage, restore the images using state-of-the-art GANs. Thanks to your efforts, mobile robots will be able to operate even in the toughest conditions. You will have a chance to work with us hand-in-hand on our computer vision and AI algorithms and to deploy your approach on commercial robotic platforms, not just GPUs in the lab. By the end of the project, you will have developed a great amount of experience related to computer vision, neural networks, GANs and embedded GPU/tensor units.
**What We Offer**
- Possibility to contribute to ongoing research in the exciting and quickly developing field of deep computer vision - Work with a team of enthusiastic roboticists and researchers in a Zurich based robotics startup in collaboration with the Autonomous Systems Lab, one of the largest robotic labs in the world. - Possibility to deploy your algorithms to different robotic platforms and highly-valued hands-on experience.
- Make yourself familiar with the current state-of-the-art image classification and restoration methods and Sevensense computer vision algorithms.
- Build upon the state of the art by developing your own ideas and your supervisor's input.
- Design, train and deploy a real-time capable solution for camera anomaly detection. Collect datasets and demonstrate the performance of your approach.
- Follow up with developing an approach that would not only detect anomalies, but also correct them by removing defects from the image.
- Design and conduct experiments with a mobile robot in challenging conditions to evaluate the approach you developed.
- Make yourself familiar with the current state-of-the-art image classification and restoration methods and Sevensense computer vision algorithms. - Build upon the state of the art by developing your own ideas and your supervisor's input. - Design, train and deploy a real-time capable solution for camera anomaly detection. Collect datasets and demonstrate the performance of your approach. - Follow up with developing an approach that would not only detect anomalies, but also correct them by removing defects from the image. - Design and conduct experiments with a mobile robot in challenging conditions to evaluate the approach you developed.
- Strong self-motivation and curiosity for solving challenging robotic problems
- Previous experience in computer vision and deep learning
- Excellent programming skills, ideally in Python; C++ is a plus
- Experience with Linux, ROS, PyTorch/Tensorflow
- Strong self-motivation and curiosity for solving challenging robotic problems - Previous experience in computer vision and deep learning - Excellent programming skills, ideally in Python; C++ is a plus - Experience with Linux, ROS, PyTorch/Tensorflow
If you are interested, please send your transcripts and CV to Marcin Dymczyk (marcin.dymczyk@sevensense.ch), Andrei Cramariuc (andrei.cramariuc@mavt.ethz.ch) and Renaud Dubé (renaud.dube@sevensense.ch).
If you are interested, please send your transcripts and CV to Marcin Dymczyk (marcin.dymczyk@sevensense.ch), Andrei Cramariuc (andrei.cramariuc@mavt.ethz.ch) and Renaud Dubé (renaud.dube@sevensense.ch).