Friday, June 15, 2018
11:00AM US Eastern Time
Speaker: Nikki Aldeborgh, Digital Globe
This webinar will cover techniques that we’ve developed at DigitalGlobe to extract information from very high resolution satellite imagery using deep learning. I will discuss feature classification for diverse use cases such as settlement detection and property attribution, as well as semantic segmentation for the delineation of man-made features such as buildings. I will also describe the important challenges that are associated with building and deploying deep learning algorithms at the scale required by satellite imagery analytics applications, and the potential ways forward that we’ve identified.
Nikki Aldeborgh is a data scientist on the Analytics and Applied Artificial Intelligence team at DigitalGlobe, where she develops algorithms to extract information from DigitalGlobe imagery, as well as the infrastructure for deploying these algorithms at scale on GBDX. She holds a Bachelor’s degree in Biochemistry from Vassar College, after which she worked on medical image processing at Memorial Sloan Kettering Cancer Center for several years. She has since honed her machine learning skills at an intensive data science course, which led her to her current position at DigitalGlobe.