Abstract: Developing State-of-the-art Computer Vision Solutions for Federal Applications
Speaker: Dr. Michael Segala, Ph.D.
Originally Prepared For: NVIDIA GTC20
Reducing manual processes and automatically identifying critical features in images and videos is transforming the industry and the federal and public sectors. Utilizing state-of-the-art computer vision has become the vital backbone in examining large volumes of data in real-time for inspection, threat detection, and real-time monitoring. Programs like this face complex, significant challenges across technical and operational objectives, including validating machine and deep learning models, optimizing their performance long-term, and integrating the output into the sensitive workflows and production pipelines. We’ll demonstrate developing end-to-end AI pipelines for computer vision applications such as satellite imaging, automated drone inspection, and defect detection. We’ll discuss the strategy, engineering, deep learning-based detection and localization algorithms, and end-user implementation.