Our computer vision researchers work on a variety of interesting projects with tangible value to business and society.
Transportation: Researchers are using computer vision to decrease traffic congestion, improve air quality and reduce fuel consumption. A number of projects help collect tolls on highways, provide intelligent parking management and detect improper use of highway express lanes. Computer vision and machine learning techniques are being used to recognize license plates and vehicle types, to provide driver assistance, to identify speeding vehicles, reckless driving behavior, and other traffic anomalies. We’ve also developed software that can rapidly search through large video data bases to help officials quickly search for missing persons or suspects.
Automating document workflows: Smart computer vision techniques are being developed to automate mobile capture of business documents with enhanced user experience, and reduced time and effort in customer transactions. Projects include intelligent mobile document capture, real-time mobile capture assistance, computational photography and mobile document recognition.
Healthcare: We are exploring techniques to remotely monitor heart rhythms, heart rate and other vital signs. This is accomplished by detecting skin color changes from camera images and using signal-processing algorithms to estimate health vitals. The development of contactless monitoring could significantly reduce healthcare costs while preventing the risk of infection and patient discomfort. We also are investigating the use of video and wearable sensors to assist patients and doctors with medical procedures.
Communication and marketing: We are developing automatic ways to measure the aesthetic quality of an image to help graphic art designers quickly find the image they need that will make the highest impact on their work.
Retail: Computer vision is being used to extract a variety of analytics about people's movements and behaviors in retail stores in order to improve overall customer experience. Projects include optimizing order processing in drive-through merges, measuring customer experience time from video footage, and video-based monitoring of parking lot statistics.
Education: After asking teachers how technology could help in the classroom, researchers created the Xerox Ignite™ Educator Support System, a one-of-a-kind workflow and software solution that pushes hand-marked student work (on paper or tablets) into the digital analytics domain. Explorations also are underway on the use of video analytics in educational settings.