PhD Position in Forensic Ico-Biometrics

Project

In this project, we wish to investigate how novel machine learning approaches, especially deep learning (DL), can be applied, adapted and evaluated for biometric analysis through characterization of anatomical shape and shape variation over time and between subjects and groups of subjects from medical imaging data in different forensic applications. Given the limited amount of training data, we aim at improving the performance and usability of DL for forensic image analysis by incorporating application-specific knowledge and by integrating DL with other computational strategies for shape analysis. Our main focus is on methodological contributions towards solutions for the forensic applications, rather than on improving DL or machine learning in general in a generic way. Given our prior expertise in image-based age assessment, this will be the benchmark application. However, the determination of other biometric features, such as gender and origin, from forensic imaging data, will be studied as well.

Profile

Master degree in software engineering, computer science, electrical engineering, informatics or any other degree with a strong focus on computation.

Have a good background in machine learning

Have a good experience in relevant programming for learning purposes, e.g. python

Have an interest in computer vision and artificial intelligence in general

Good skills in oral and written English

Be interested in an intensive research experience of 4 years.

Offer

We offer a PhD position in a multidisciplinary research environment that is at the forefront of computational forensic investigations. You will be responsible for developing, implementing, and testing novel image-based applications in forensics. You will be given an insight into the interesting world of forensics and crime scene investigations(CSI) and will be given the opportunity to acquire a solid background in machine learning and image analysis.

Interested?

For more information please contact Prof. dr. ir. Dirk Vandermeulen, tel.: +32 16 32 90 39, mail: dirk.vandermeulen@kuleuven.be or Prof. dr. Peter Claes, mail: peter.claes@kuleuven.be