Lawrence Livermore National Laboratory



Technical Focus Areas: Computer Vision and Video Analytics
Image and video processing, along with statistical modeling, enable applications such as persistent surveillance

3-D Overhead Reconstruction

Our Capabilities

Livermore is both leveraging and advancing state-of-the art techniques in computer vision and video analytics to address problems of national interest. Novel approaches to processing wide-area motion imagery have resulted in superior data products for analyst exploitation and archival storage. In addition, techniques such as object recognition, change detection, and tracking have demonstrated the potential to improve analyst efficiency as well as provide new capabilities. Cutting-edge machine learning and video analysis methods are also being explored to enable large-scale search, retrieval, and indexing. This work focuses on applying and extending state-of-the-art techniques in unsupervised feature learning on large data sets in ways that use all available information: audio, imagery, motion, semantic information, and tags to aid in building large-scale learners that facilitate flexible classification and query strategies.

Recent Publications

  • Sakla, Wesam, Goran Konjevod, and T. Nathan Mundhenk. "Deep Multi-modal Vehicle Detection in Aerial ISR Imagery." Applications of Computer Vision (WACV), 2017 IEEE Winter Conference on. IEEE, 2017.
  • Mundhenk, T. Nathan, Laura M. Kegelmeyer, and Scott K. Trummer. "Deep learning for evaluating difficult-to-detect incomplete repairs of high fluence laser optics at the National Ignition Facility." The International Conference on Quality Control by Artificial Vision 2017. International Society for Optics and Photonics, 2017.
  • Mundhenk, T. Nathan, Goran Konjevod, Wesam A. Sakla, and Kofi Boakye. "A Large Contextual Dataset for Classification, Detection and Counting of Cars with Deep Learning." In European Conference on Computer Vision, pp. 785-800. Springer International Publishing, 2016.
  • Kwon, Youngwook P., Hyojin Kim, Goran Konjevod, and Sara McMains. "Dude (Duality descriptor): A robust descriptor for disparate images using line segment duality." In Image Processing (ICIP), 2016 IEEE International Conference on, pp. 310-314. IEEE, 2016.
  • Grathwohl, Will, and Aaron Wilson. "Disentangling Space and Time in Video with Hierarchical Variational Auto-encoders." arXiv preprint arXiv:1612.04440 (2016).
  • Sawada, Jun, et al. "Truenorth ecosystem for brain-inspired computing: scalable systems, software, and applications." High Performance Computing, Networking, Storage and Analysis, SC16: International Conference for IEEE, 2016.
  • Boakye, K., P. Kidwell, G. Konjevod, and J. Lenderman. Literature review for vehicle correspondence and network modeling and analysis. No. LLNL-TR-680306. Lawrence Livermore National Laboratory (LLNL), Livermore, CA, 2015.

Recent CASIS Workshop Proceedings

Contact a Subject Matter Expert

Alan Kaplan, (925) 423 0161, kaplan7@llnl.gov
Computer Vision Group Leader

In the Spotlight —
Persistics Persistent Surveillance

The Challenge

Wide-area motion imagery for persistent aerial surveillance generates enormous amounts of video data — many terabytes per minute — overwhelming both analysts and storage systems.

The Results

Persistics Imagery

Until now, we had no practical way to store that much data. With Persistics, we have an innovative method to compress the equivalent of thousands of hard drives to just a few drives.

Holger Jones, LLNL

Learn More »