The Learning and Machine Perception (LAMP) team at the Computer Vision Center conducts fundamental research and technology transfer in the field of machine learning for semantic understanding of visual data. The group works with a wide variety of visual data sources: from multispectral, medical imagery and consumer camera images, to live webcam streams and video data. The returning objective is the design of efficient and accurate algorithms for the automatic extraction of semantic information from visual media.
Two papers will be presented at the NIPS workshop on Adversarial Training. “Ensembles of Generative Adversarial Networks” investigates the usage of ensembles for GANs, and shows that they can significantly improve results. “Invertible Conditional GANs for image editing” adds a encoder to a conditional GAN, thereby allowing to do semantic editing of photos. The code is available here.
Congratulations to all our master students that defended their thesis last 16th of September during the 3rd Annual Catalan Meeting on Computer Vision!
- Olaia Artieda Aguirre
- Sergi Canyameres Masip
- Xialei Liu
- Arcadi Llanza Carmona
- Guim Perernau Guirao (Best Thesis Award)
We also received the Best Poster Award for our poster based on the ‘Hierarchical Part Detection with Neural Networks‘.
Our paper “Does Multimodality Help Human and Machine for Translation and Image Captioning?” has been accepted on the ACL 2016 First Conference on Machine Translation. Check some results here.
Two of our Ph.D. students, Laura López and Marc Masana, participated in the Accenture Digital Datathon. The competition consisted on generating a model for traffic accident prediction in the city of Barcelona. Our student’s team won the Business Track by unanimous decision of the jury.
Our paper ‘On-the-fly Network Pruning for Object Detection‘ has been accepted at ICLR 2016 on the Workshop Track.
Also, the paper ‘Hierarchical Part Detection with Deep Neural Networks‘ has been accepted at the ICIP 2016.
The work ‘Fitnets: Hints for thin Deep Nets’ will be presented at the ICLR 2015. The paper proposes a framework to compress wide and deep networks into thin and deep ones.