Today most systems designed for defense, observation from space, transportation, homeland security or critical infrastructure protection are embedding image or video in order to help in decision making processes.
The aim of this presentation is to highlight the needs and expectations for image and video processing in these domains.
Starting from a survey of the main applications, we identify the driving forces (e.g. evolution of technologies, decreasing cost of components, new sensors, new customers needs, similarities between some civilian and military applications, …) which could lead to potential evolutions in years to come. In particular, image or video analyses are still perceived in most applications as self managed component of an autonomous system while they will become more and more closely inter-related with other system components (e.g. human in the loop, optical and mechanical parts, GIS, Internet of Things, other sensors, …) in order to boost the global performances of critical systems. For instance, in the space domain, the high level of user application requirements leads to the necessity to make that different types of sensor collaborate, either after acquisition (data assimilated into models) or directly during acquisition (collaboration at mission level). In this context, the satellite is no longer a closed system but has to be designed as a component of a complex system.
Finally we conclude by listing the main perceived technological barriers which are yet to overcome to achieve the considerable change this will introduce.
Graduated from Ecole Polytechnique (2002) Jean-François GOUDOU has a PhD in image processing from Telecom ParisTech / Sagem Defense & Security for detection in naval infrared search & track, using spatio-temporal filters and super-resolution. He is leading several projects in Thales Theresis laboratory. Since 2009 he has been in charge of the demonstrators of the VisionLab, the joint corporate lab setup by THALES and CEA for video analysis. He also led several research projects in the topics of video-surveillance and algorithmic evaluation, especially the ITEA2 ViCoMo project French consortium (Video Context Modelling) and the French Quasper R&D and Methodeo projects.
His current main domains of interest are human vision, neurosciences and human-inspired machine learning technologies.
Olivier Koch runs the Image Processing department within THALES Optronique SA, responsible for the design and the development of image-processing algorithms for a set of optronic products (cameras, sighting systems, pods) embedded in ground, maritime, and airborne defense systems. He started his career as real-time software engineer for General Electric Healthcare. He graduated from ENSTA-Paristech (2002) and has a Ph.D. in computing and artificial intelligence from the Massachusetts Institute of Technology (2010).
Adrien GIRARD is a Technical Leader at Thales Alenia Space in charge of radiometric and geometric performances of space imaging systems. His work encompasses analysis of optical image chains, end-to-end simulations of satellite imagery, image quality improvement techniques, and on-orbit sensors calibration and characterization. He has been a research assistant in image processing at Trinity College Dublin and graduated from TELECOM ParisTech in 2008.
Jean-Yves Dufour received a Ph.D. in image processing and pattern recognition from the University of Caen in 1988. He then joined THALES Optronique SA as a research engineer in image processing before leading the development of image-processing-based functions for optronic products and running the image-processing department. Early 2009, he joined THALES Services as technical manager of the joint corporate lab setup by THALES and CEA for video analysis. His current main technical domains of interest are crowd analysis and automatic detection of “abnormal” activities in a scene. In 2012, he edited a collective book dedicated to Video Analytics (“Intelligent Video Surveillance Systems,” ISTE/WILEY), written by principal French leaders in the domain.
Abstract: This workshop presents a collection of MATLAB techniques useful for image-processing practitioners. The techniques include algorithm-implementation patterns, methods for visualizing results, creating interactive image apps, and sharing results with colleagues. Even experienced MATLAB users will come away with something new and useful.
Steve Eddins, an electrical engineer turned software developer, has been at MathWorks since 1993, developing MATLAB and image-processing products. He led development of the Image Processing Toolbox for 15 years. During that time, he designed and implemented a broad spectrum of image-processing capabilities, including filtering and filter design, transforms, mathematical morphology, geometric transformations, image analysis, segmentation, color science, visualization, and image file formats. He also created the 2nd-generation MATLAB performance profiler and the MATLAB xUnit Test Framework. Today, Dr. Eddins is a senior MATLAB designer for language evolution and for the overall collection of MATLAB functions and types. He coaches MathWorks development teams on the design of programming interfaces intended for use by engineers and scientists. Before joining MathWorks, Dr. Eddins was on the faculty of the Electrical Engineering and Computer Science Department at the University of Illinois at Chicago. There, he taught graduate and senior classes in digital image processing, computer vision, pattern recognition, and filter design, and he performed research in image compression. Dr. Eddins co-authored the book Digital Image Processing Using MATLAB and writes regularly about image processing and MATLAB at http://blogs.mathworks.com/steve. Dr. Eddins holds a B.E.E. (1986) and a Ph.D. (1990), both in electrical engineering, from the Georgia Institute of Technology. He is a senior member of the IEEE.
Abstract: While well known for search, Google has now grown to generate significant impact in the media-processing space. Recruitment in media/imaging and vision has been growing for some time. This workshop features three Googlers giving 20-min snapshots of Imaging and Video DSP technology currently being explored by Google/YouTube. We highlight key developments and expose some of the underbelly of technology research and development in YouTube, Chrome, and Glass. Specifically, the speakers will cover the following topics:
Debargha Mukerjee “VP9 – Adoption and Beyond”: Google finalized a next-generation open-source royalty-free video codec called VP9 as part of the libvpx repository of the WebM project (http://www.webmproject.org/) in June of 2013. This talk will provide the attendees an overview of the more recent developments in the VP9 eco-system since then, including incorporation of high-color and high-bit-depth profiles, as well as the current state of adoption of VP9—primarily in YouTube. Furthermore, the workshop will provide a technical overview of the new experimental coding tools that are being explored on a divergent branch of the libvpx repository—called the playground—that will hopefully mature into a future-generation codec. The workshop will also provide pointers on how interested parties can be involved in the project.
Anil Kokaram “VideoDSP at YouTube”: YouTube is known for its size: 100 hours of video uploaded per minute, millions of views per day, hundreds of supported video formats, worldwide broadcasting. This talk considers the enabling technology behind those numbers and highlights the challenges faced by the teams deploying robust video-DSP tools in that space. We also touch on interactions between YouTube and Chrome especially with regard to 4K broadcasting.
Peyman Milanfar “Computational Imaging at Google”: Fancy cameras and high-quality imaging used to be the exclusive domain of professional photographers. Times have changed, but even as recently as a decade ago, consumer cameras were solitary pieces of hardware and glass; disconnected gadgets with little brains, and no software. But now, everyone owns a smartphone with a powerful processor, and every smartphone has a camera. These mobile cameras are simple, costing only a few dollars per unit. And on their own, they are no competition for their more expensive cousins. But coupled with the processing power native to the devices in which they sit, they are so effective that much of the low-end point-and-shoot camera market has already been decimated by mobile photography. Computational imaging is the art, science, and engineering of producing a great shot (moving or still) from inferior sensors. It does so by changing the rules of image capture—recording information in space, time, and across other degrees of freedom—while relying heavily on post-processing to produce a final result. As an example, Google Glass is the first head-mounted, burst-based mobile camera on the market. Coupled with the ubiquity of mobile devices, powerful algorithms and open platforms for imaging will inevitably lead to an explosion of technical and economic activity, as was the case with other types of mobile applications.
Debargha Mukherjee received the Ph.D. in Electrical and Computer Engineering from the University of California Santa Barbara in 1999. Between 1999 and 2010, he was at Hewlett Packard Laboratories conducting research on video and image coding and processing. Since 2010, he has been with Google where he is currently involved with open-source video-codec development. Dr. Mukherjee has authored or co-authored more than 75 papers and holds more than 20 US patents. He was a student-paper award winner at IEEE ICIP in 1998 and, in 2012, he was a keynote speaker at both the SPIE Stereoscopic Displays and Applications Conference and the CVPR 3D Cinematography Workshop. He currently serves as an associate editor of the IEEE Transactions on Image Processing and the SPIE Journal of Electronic Imaging.
Anil Kokaram is a Tech Lead in the Transcoding Group at YouTube/Google, leading a group responsible for video quality. He is also a professor at Trinity College Dublin, Ireland, and continues to supervise a small number of students at http://www.sigmedia.tv in the Department of Electrical Engineering. His main expertise is in the broad areas of DSP for video processing, Bayesian inference, and motion estimation. In 2007, he was awarded a Science and Engineering Academy Award for his work in video processing for post-production applications. He founded a company (GreenParrotPictures) producing video-enhancement software that was acquired by Google in 2011. He is currently an associate editor of the IEEE Transactions on Circuits and Systems for Video Technology.
Peyman Milanfar has been involved with the Glass project at Google and now leads a team within Google Research focussed on imaging. He was Associate Dean for research and graduate studies at the University of California Santa Cruz (UCSC) from 2010 to 2012. He received the Ph.D. in Electrical Engineering and Computer Science from MIT. Prior to UCSC, he was at SRI, and a Consulting Professor of Computer Science at Stanford. In 2005, he founded MotionDSP, which brought state-of-art video enhancement to market. His interests are in statistical signal, image, and video processing; computational photography; and machine vision. He is a Fellow of the IEEE.