Description: Writer identification is the task of determining the author of a sample handwriting from a set of writers. This problem has attracted a great deal of attention due to the large number of potential applications, such as personalized handwriting recognition, automatic forensic document examination, writer retrieval, and classification of ancient manuscripts. However extracting discriminant features still an open problem.
There is an immediate opening for a funded full-time PhD position at ÉTS (École de technologie supérieure) in the area of handwritten document authentication. The candidate will work under the supervision of professors Robert Sabourin (ÉTS) and Luiz Oliveira (UFPR, Curitiba, Brazil).
Preferred candidates should have a solid background in deep learning in general, and more specifically in CNN-based feature learning and computer vision.
Application Instructions: Applicants should have strong academic record in computer science or in electrical engineering. A Master degree with expertise in Machine learning, Pattern Recognition and Computer vision. Requirements: Experience with Caffe, Torch, Keras and/or Theano frameworks is a must. Proficiency in C, C++, Python and Matlab. Experience with CUDA and OpenCV.
Contact: Prof Robert Sabourin (robert DOT sabourin AT etsmtl DOT ca)