Computer Vision News - July 2016

No doubt, our readers remember the Face2Face research paper which we have reviewed in our May issue . It happens that Matthias Nießner and his friends were at CVPR and gave a presentation which was probably one of the greatest moment of the Conference. They were brave enough to offer us an outstanding live demo at the CVPR orals. Don't miss a short but very meaningful video below! 8 Presentations Face2Face BEST OF CVPR DAILY This work constructed a large dataset of casual walkers from egocentric videos with weather and location information. Without requiring manual annotations, this project proposed a self-supervised pre-training model by leveraging discretized contextual information (geo-location and weather) as weak labels to learn features suitable to facial attributes. By learning with diverse contextual information, the framework could be also applied to other high-level analysis tasks. The full paper is here . Jing Wang presented the Walk and Learn model Walk and Learn