***For Immediate Release***

Canesta Introduces Electronic Perception Technology To The Machine Vision Community

Low Cost, Advanced Technology Resolves Many Dilemmas, Such as "Is It Big or Is It Close?"

BOSTON, MA – May 4, 2004 – Canesta's low-cost electronic perception technology is being rolled out in a new area for the first time today — for industrial "machine vision" applications. Machine vision refers to the use of cameras and image processing software with robots or other types of machines to enable them to automatically locate, identify, measure or otherwise inspect objects, particularly in industrial settings. Canesta's breakthrough technology utilizes tiny CMOS-based sensors that are capable of "seeing" in three dimensions, which enables them to determine the size, shape, and exact distance of nearby objects, without the use of multiple cameras or sophisticated material handling equipment.

"This is an important milestone for the machine vision community," said Jim Spare, Canesta's vice president of product marketing. "For the first time, a small, single-chip imaging solution can replace complicated scanning systems, stereoscopic setups, or expensive processing using complicated image processing software." According to Spare, the low cost and small size of the sensor chip — about the size of a pencil eraser — plus the extraordinary amount of processing that it eliminates, means that virtually any type of equipment or device can be equipped to "see the world around itself" in true, 3-dimensions.

Conventional imaging technology used in machine vision applications depends upon familiar camera technology that renders a three dimensional world into flat, two-dimensional images that are, essentially, arrays of colored (or gray) dots, or 'pixels.' To reconstruct a 3-dimensional view — which is important for applications that require understanding of the spatial locations of objects in the nearby environment — is a profoundly difficult problem, one that has given rise to an entire sub-discipline in the field of machine vision. Canesta solves this reconstruction problem by eliminating it altogether. Instead, a single Canesta sensor can acquire — rather than derive — primary 3-D information, and output this data at over 30 frames per second to provide a real-time, moving topographical or "relief" view a machine's surroundings. "We also see the world in arrays of pixels," said Spare, "but we know exactly how far each one is from our sensor, and we report it to the machine."

One problem that is immediately solved is that of perspective. "It is a challenge in machine vision to determine if an object is large or simply close by," explained Spare. "If you take a photograph of a friend with a mountain in the distance, your friend's head will appear larger than the mountain. You know the difference, but a machine does not, without 3-dimensional — or distance — information. But although the mountain is a little out of range, our chip could report the exact distance to your face, as well as your nose, or eyes, or ears." Spare said that a robot with Canesta's 3-D sensor chips could easily be designed to fit you with sunglasses.

Applications for Canesta's 3-D sensor chips are numerous. In robotics, the availability of primary 3-D data could reduce error, complexity, and response delay. Manufacturing, warehousing or security applications would be able to locate and sort objects by size and shape. Cargo sensors could monitor cargo load or package size based upon volume. And automatic doors could be improved such that they only opened when individuals approached them, rather than each time someone walked by. "There's no limit to the possibilities," said Spare.

Canesta's electronic perception technology will be showcased at the Vision Show east, running here from May 4th through 6th. The Vision Show East, sponsored by the Automated Imaging Association, attracts experts from the fields of machine vision, robotics, materials handling, and numerous other automation disciplines. The Show appeals to the developers and users of machine vision solutions for a wide variety of industries, including automotive, electronics, food & beverage, medical/pharmaceutical devices, printing, semiconductors, and wood products.

Canesta's electronic perception technology is available to original equipment manufacturers serving these fields, and will be demonstrated at the show. Prototyping kits for further development will be introduced during the summer.

About Canesta

Canesta is the inventor of a revolutionary, low-cost electronic perception technology that enables machines and ordinary electronic devices to perceive and react to nearby objects or individuals in real time.

When sight-enabled with Canesta's unique electronic perception chips and software, consumer, automotive, industrial, military, and medical products can gain functionality and ease of use not possible in an era when such devices were blind.

Canesta believes future applications of electronic perception technology are virtually as broad as the imagination. They may include intelligent automobile airbag systems that can sense the size and position of an occupant to control deployment and avoid injury, a low-false-alarm security system that could detect the difference between an intruder and normal activity, such as a pet moving or child visiting the bathroom at night, or robotic tools that can successfully operate in a dynamic, rather than static environment.

Canesta was founded in April 1999, and is located in San Jose, CA. The company has filed in excess of forty patents, 12 of which have been granted so far. Investment to date exceeds $36 million, from Apax Partners, Carlyle Venture Partners, Intel Capital, JP Morgan Partners, Korea Global IT Fund (KGIF), TechFund Capital, Thales Corporate Ventures, and Venrock Associates.

-30-