DataScience@UB designs and prototype a machine that includes an intelligent vision system based on semantic segmentation through convolutional neural networks to automate the classification of leather.
Application of deep learning techniques in the detection of changes and generation of maps of land covering of Catalonia (MCSC).
Within the framework of the Cartographic Plan of Catalonia, the Cartographic and Geological Institute of Catalonia is working on the elaboration of the technical specifications of the land covers Map of Catalonia. The ICGC has commissioned us a study to explore the viability of the application of deep learning techniques in the generation of maps.
Up to now, land cover maps of the territory have been manually worked on from orthophotographic images. The comparison of the covers between the different temporal periods in which images were taken was done manually as well. The DataScience@UB group is developing a model in order to include image segmentation processes based on deep learning techniques for the automatic construction of maps of the land covers of the territory.
Dispose of updated maps at shorter time intervals favoring a better interpretation of the territory for territorial and urban planning.
Thanks to the emerging technologies of e-Learning, the analysis of data offers professionals in the education sector, advantages with the ability to impact and revolutionize in the future the way we analyze and evaluate the learning experience.
A company in the online sector has generated a large volume of data that provides a history to be able to extract information from clients and develop a predictive model based on the behavior of users. This behavior include the navigation through the platform, the interaction with others channels, messages to the community, etc. The ultimate goal is to provide a better customer service by recommending products based on the specific time of the customer journey.