Sight is one of the most fascinating of the human senses. Since the dawn of computers, humans have been trying to build machines that can mimic human visualization and interpretation. At GE Research, fundamental and applied computer vision technology research is performed using a plethora of tools and techniques to tackle various problems. From classical image processing and computer vision techniques that have proven their effectiveness over the years to emerging AI technologies such as Deep Learning—we use the tools that befit the job.
How is GE Research shaping the future of computer vision and image processing?
Computer vision has been in existence for more than three decades at GE Research, building on a strong legacy of groundbreaking work in visualization and image processing. From the genesis of object-oriented programming techniques to industry-standard software such as ITK, VTK and 3D Slicer, GE Research has made significant contributions to the development of visualization, image analysis and computer vision for decades.
GE Research performs interdisciplinary computer vision research, leveraging the domain expertise of people who know the physics of the systems, researchers who are adept at developing robust software architectures to embellish our computer vision algorithms and roboticists who can build intricate robots that can help us gain access to data.
Computer vision applications
GE Research works in close coordination with GE's business units and with several government scientific organizations, such as NIH, IARPA and DARPA, in the realm of computer vision applications. Topics of research span a wide spectrum of application including biomedical image analysis, video analytics and inspections systems.
Biomedical image analysis
GE Research builds computer vision systems for biomedical image analysis that can interpret medical scans (CT, MRI, Ultrasound, etc.) or even microscopic images of cellular structures, faster and more quantitatively than humans.
Inspection systems is an area of ever increasing importance in the computer vision industry, where we inspect newly made parts or parts that have come in for servicing for manufacturing or operational defects such as pits, dents, cracks, nicks and other signs of fatigue.
Video Analytics is another exciting computer vision domain where we build intricate systems that analyze video feeds in a closed loop control that can actively or passively interact with the subject being watched to understand group dynamics, behavior, intent and develop situational awareness.