PhD Candidate in Computer Vision

NTNU is a broad-based university with a technical-scientific profile and a focus in professional education. The university is located in three cities with headquarters in Trondheim.

At NTNU, 9,000 employees and 42,000 students work to create knowledge for a better world.

You will find more information about working at NTNU and the application process here.

…

(Video unable to load from YouTube. Accept cookie and refresh page to watch video, or click here to open video)

About the position

We have a vacancy for a PhD candidate in Computer Vision, at the Department of Computer Science, funded by the NTNU Digital programme.

For a position as a PhD Candidate, the goal is a completed doctoral education up to an obtained doctoral degree.

In computer vision, the measurement of object similarity is the key ingredient for operations such as object recognition, detection and retrieval. As objects are commonly represented by proxies called object descriptors, similarity evaluation is based on designing robust and accurate object descriptors and good distance functions between such descriptors. State-of-the-art 3D object descriptors are mainly concerned with object shape. We argue that in order to globally address the above operations, one must consider both shape and spectral information conjointly, and in order to do so new descriptors and distance functions are needed. The challenging fundamental computer science research required for this will subsequently enable applications in a broad range of fields, such as cultural heritage (e.g. reassembly from fragments, erosion measurement), medicine (e.g. pathological object detection), and autonomous vehicles (object detection and tracking for navigation).

We believe that by integrating the shape and spectral descriptors, building on our state-of-the-art previous work in each of them, we can achieve significant progress in the following key objectives:

• Develop novel spectral + shape descriptors and distance functions.
• Validate these using appropriate data sets.
• Apply them in areas such as estimating the degradation of cultural heritage objects, and reassembling cultural heritage objects from fragments.

Further specification of the topic and project plan will be done in collaboration between the selected candidate and the PhD supervisors, Prof. Jon Y. Hardeberg and Prof. Theoharis Theoharis. 

The main place of work for the position is the Colourlab in Gjøvik, but significant periods of time will also be spent at the Visual Computing Lab in Trondheim.

Your immediate leader is Head of Department.