Title: Deep learning for an easier extraction of the information from remote sensing images
Summary: Recently, deep learning approaches have shown a strong potential for many computer vision tasks. In particular for remote sensing, these approaches give new opportunities and new challenges. During this seminar, I will present a new task to simplify the extraction of information from remote sensing images. This approach aims at answering to questions formulated in natural language on a given image. I will present my work on this task, including a method for the generation of reference data and a model taking into account a specificity of remote sensing images.
Practical information
- Time: 18/06/2020, 16h
- Place: Zoom
Links
- Slides
- RSVQA page
- Article the conversation
- Project (restricted access)