»Radar in Action« Machine Learning for Radar Applications

Описание: Have you missed our live lectures? We are now publishing selected presentations of #RadarInAction on #Youtube! If you have any questions about the topic, simply asks us in the comments!


»Radar in Action« Machine Learning for Radar Applications – Small Target Detection and Robust Classification using Neural Networks

Machine Learning approaches, especially neural networks, give State-of-the-Art results in several computer vision tasks. This presentation shows two applications of neural networks to radar data. The first one is the detection of persons and drones with a surveillance radar. The second one is the classification of radar targets using convolutional neural networks. The latter part focuses on the robustness of these networks, i.e. which part of the input data is relevant for classification and how to avoid fooling.

Speaker:
M. Sc. Simon Wagner

Themes:
Small Target Detection
Classification of Radar Targets
Which part of the input image is used for classification
Methods to avoid fooling of neural networks

Contact and more info:
https://www.fhr.fraunhofer.de/en/even...


What is »Radar in Action«?

By presenting our research results, we want to show what is possible with radar and how you can use the possibilities. Unfortunately, this is hardly possible in person at the moment, so in 2020 we launched the public online lecture series »Radar in Action«. The lectures are aimed at customers, partners and interested parties from industry, politics, science and society.

In 2020, there were 20 events with an average of about 100 participants, totaling over 900 people, many of whom attended multiple times. Over 93% rated the presentations as very good or good.

The events are always held on Tuesdays from 2:00 pm to 2:30 pm CET. They include live presentations on a radar application - some including live demonstrations using experimental systems, as well as opportunities for questions and exchanges. Participation is free of charge, only registration is required.

Visit https://www.fhr.fraunhofer.de/radar-i... for more topics and dates of our online live events. The registration is free of charge.