Typically training deep neural networks require large amounts of annotated data and processing power to reach good results. It’s often beneficial to pretrain a neural network using another dataset to reduce the annotation work and computations needed for the application.
SICK, as a sensor provider, is in a unique position to get image data from many different domains. For example, depth or multispectral images. In these domains easily reusable datasets for pretraining are not readily available. Sample datasets from customers are often relatively small and/or difficult to annotate due to the unfamiliar image characteristics.
This thesis proposes to automatically pretrain a deep neural network using a very limited dataset in non-natural image domains. The aim is to generate a neural network that works better for downstream task on this or similar datasets, compared to pretraining on a larger standard dataset.
To evaluate the success of the method there are at least two applications.
One within image anomaly detection and the other within classification.
For more information, contact:
Daniel Rydström, Group Manager, firstname.lastname@example.org
Charlotte Axelsson, HR Manager, email@example.com
Welcome with your application 18th of October at the latest!
SICK is a world-leading supplier of sensors and sensor solutions for industrial applications. We are 10 000 employees in 50 countries and our head quarter is located in Freiburg, Germany. SICK in Linköping is an innovation center for Machine Vision and we are 75 engaged employees with a big interest in image processing and visualization. For more than 35 years, our team at SICK Linköping has successfully developed and delivered software for technical leading products within the field of 2D and 3D vision, as well as system solutions for i.e. robot guidance and quality control.
At SICK in Linköping we are very proud of our work to be a healthy and attractive workplace. We have been elected one of the best workplaces in Sweden according to the survey Great Place to Work 2021 and several years before that. We work actively to reduce our climate footprint and engage in various ways to contribute to society and increase diversity at our workplace.
Annonsen är avslutad och det går inte längre att ansöka.