IEEE SPS Summer School on Light Field Data Representation, Interpretation and Compression
July 28, 2018

IEEE have recently organized a summer school on Light Field Data Representation, Interpretation and Compression and invited Prof. Paolo Favaro to give a talk. His presentation covered topics in light field processing such as depth estimation, extended depth of field, super resolution, reflection separation and motion deblurring. The school took place at Mid Sweden University in Sundsvall, Sweden during May 28 - June 1, 2018.

Rank Prize Symposium
July 28, 2018

Professor Paolo Favaro and PhD student Mehdi Noroozi are invited to the Rank Prize Symposium on Computer Vision, which is to be held in Grasmere, England, from August 20-23. It will bring together 10 senior and 20 early career researchers to discuss the topic of geometry and uncertainty in deep learning for computer vision. Among the speakers are Alyosha Efros, Raquel Urtasun, Vladlen Koltun, Daniel Cremers, Paolo Favaro, Andrew Blake, Andrea Vedaldi, Andrew Zisserman and Alex Kendall.

ICVSS 2018
July 27, 2018

Prof. Paolo Favaro was invited to give a talk at the International Computer Vision Summer School (ICVSS) that took place in Sicily, Italy. The topic of this year’s school was “Computer Vision after Deep Learning”. In his talk “Beyond Supervised Learning”, Prof. Favaro presented state-of-the-art methods to extract features from data without labels, and methods to disentangle factors of variation in the data. You can read the abstract here.

Best Poster Award for Simon Jenni
July 27, 2018

Recently, PhD student Simon Jenni visited the AI Summer School P.A.I.S.S. in Grenoble and received the award for best poster for his CVPR work Self-Supervised Feature Learning by Learning to Spot Artifacts. Congratulations!

Our CVPR, ECCV and SIGGRAPH 2018 Papers
July 27, 2018

We are very happy to announce that we have a total of eight accepted papers in the conferences CVPR, ECCV and SIGGRAPH. For CVPR we have a new record of four accepted submissions:

A common theme of all of these works is “self-supervised learning”, a trending topic in Computer Vision and Deep Learning that our group is very much interested in. Furthermore, in ECCV we have

  • "Deep Bilevel Learning" by Jenni et al.,
  • "Understanding Degeneracies and Ambiguities in Attribute Transfer" by Szabó et al.,
  • and "Normalized Blind Deconvolution" by Jin et al.

We also have a SIGGRAPH work FaceShop: Deep Sketch-based Image Editing by Portenier et al. that got attention in the press. You can read the article here.