We are looking for a Postdoctoral Research Fellow in the field of 3D Shape Analysis, combining Deep Learning with ‘traditional’ geometric and statistical approaches to Computer Vision. This position is part of the project 3D Shape Analysis with Geometric Declarative Networks, supported by the Royal Society / Te Apārangi through the Marsden Fund. The project will advance 3D shape analysis, going beyond current coarse classification methods to fine-grained classification on the basis of shape alone. To achieve this, we will use recently proposed declarative network architectures to integrate collaborative representations and geometric deep learning. This combination will allow more detailed analysis of 3D objects. In particular we will be able to distinguish broadly similar objects on the basis of local shape features. Our initial application is in archaeology, specifically traditional Māori stone tool manufacture. Subtle variations in the waste flakes from this process make them an ideal target for our proposed methods.

Plus d’informations :
[Website University of Otago]