Groundbreaking Image-Analysis Technology Puts Teradyne Machine Vision
on Leading-edge
Teradyne's (NYSE: TER) Assembly Test Division, a global leader in electronics testing and production solutions, announced today that the inventor of a significant Teradyne inspection technology--Dr. Pamela Lipson and her company Imagen--have been recognized as one of an elite group of 40 "Technology Pioneers for 2003" by the World Economic Forum in Geneva, Switzerland. Lipson and Imagen were recognized for the development of innovative technologies for recognition and analysis of complex visual and image-based information.
Teradyne has licensed this innovative technology and incorporated it as the fundamental image-analysis software technology in the Optima(TM) 7200 OPT machine vision system, which will be formally introduced at the APEX trade show, a major electronics-manufacturing industry event, in Anaheim, Calif. on March 31.
The leading-edge technology was initially developed by Dr. Lipson and her colleagues at the Massachusetts Institute of Technology's Artificial Intelligence Lab, and was refined by her company Imagen Inc. In 1997, the technology attracted the interest of Teradyne, and was developed for application in the field of printed circuit board inspection. It is now the basis of Teradyne's newest generation of Optima machine vision systems. Several of the new Optima 7200 OPT systems, jointly developed by an integrated team of Imagen and Teradyne engineers, are already in use by electronics manufacturers in Asia and North America.
"Pam's recognition by the World Economic Forum is quite an honor," said John Casey, President of Teradyne's Assembly Test Division. "We are very happy to see her recognized for her substantial achievements in developing this significant technology, and are excited to have been able to leverage it to introduce a major advancement in the science of PCB vision inspection."
The Imagen technology delivers intelligent image analysis and object recognition capability to `test' visual objects collected in a composite image of a manufactured circuit board. The images are analyzed for defects and for manufacturing process trends that indicate future defects. The recognition technology allows the system to test the specific objects in the PCB appearance without lengthy programming or training efforts. Dr. Lipson's technology also delivers built-in tolerance to normal object appearance variation (for instance, from supplier changes), to accelerate inspection times.
"Pam and her team developed an innovative approach to visual recognition that has revolutionized the domain of industrial machine vision," noted Eric Grimson, Bernard Gordon Professor, Department of Electrical Engineering & Computer Systems at MIT. "Exhibiting a tolerance to environmental effects that challenge that of a human, while recognizing minute nuances in part positioning and trends across time, these methods provide a level of remarkably robust performance at incredible speeds. The potential for applying these methods in domains beyond industrial machine vision is enormous."
Feedback from early users of the Optima 7200 OPT system has been very positive. Teradyne will introduce this new technology in the Optima 7200 OPT machine vision systems in booth # 2050 at the APEX 2003 trade show in Anaheim, Calif. from March 31 - April 2.
About Imagen Inc.
A young spin-off company from the MIT Artificial Intelligence Laboratory, Imagen Inc. specializes in automated visual recognition and pictorial/multimedia search systems. At the heart of Imagen is a groundbreaking technology for image and video analysis that was developed by the founders at MIT. For more information visit www.imagen-inc.com.
About Teradyne
Teradyne (NYSE: TER) manufactures Automatic Test Equipment and interconnection systems that deliver competitive advantage to the world's leading semiconductor, electronics, automotive and network systems companies. Teradyne had sales of $1.22 billion in 2002 and employs about 7200 people worldwide. For more information visit www.teradyne.com.
“Exhibiting a tolerance to environmental effects that challenge that of a human, while recognizing minute nuances in part positioning and trends across time, these methods provide a level of remarkably robust performance at incredible speeds. The potential for applying these methods in domains beyond industrial machine vision is enormous.”

