Nanostructured objects represent increasingly important products of nanotechnologies having a wide range of industrial applications in textiles, medical imaging, automotive, semiconductors etc. Such objects contain structures smaller than 100 nm in at least one dimension. Estimation of shape and size of such objects is important in the context of safe by design. The project aims to develop objective and automatic approaches of image segmentation and classification of nanoparticles. The aim of the present project will be twofold:
- to train the system classifier using images of nanoparticles
- to design a database for metadata storage and retrieval
The student will benefit from the existing expertise in image and morphological analysis in the group and is expected to develop an approach based on a combination of machine learning and image analysis approaches. The student will extend an already existing platform developed in collaboration with a team in the Zuse Institute Berlin, called Active Segmentation. The platform is based on ImageJ and allows researchers, who are not experts in image segmentation, to use advanced filtering and machine-learning techniques for object segmentation and classification. The users train the system on examples using only their domain-specific knowledge of the subject.
Experience with Java
Desired skills: experience with machine learning platforms.
Type of project: Combination of internship and thesis, Thesis
Duration: 6-8 months
Required degree: Master of Science
Required background: Computer Science, Nanoscience & Nanotechnology
Supervising scientist(s): For further information or for application, please contact: Dimiter Prodanov (Dimiter.Prodanov@imec.be)