IDF Fall 2010 Day 0 - Intel R&D Demos


Intel's New Approach

Intel has been taking a new direction when it comes to the development of their future products. Through the eyes of an anthropologist like Dr. Genevieve Bell, who is the director of Interaction & Experience Research at Intel Labs, understanding user behavior and their needs are top priority at Intel - that is if they want to create future products that remain useful and relevant.

According to Bell, Intel needs a new perspective in the way they create chips and systems to deliver the required user experience. In the past, the PC has always been modified and tweaked to make applications work for the user, but at Intel Labs, they are first exploring how the user's experience could be translated and built into the silicon. This means that compute performance can be enhanced at the chip level, with each feature made for the user in mind. Intel hopes that this will translate into products that will deliver the required performance while addressing the various needs of the future user.

At Intel Labs, they are changing how they design their products, and it's no longer just about building new technologies and then search for problems that it could solve. Instead, by combining engineering and other disciplines like anthropology, engineers are challenged to find solutions to everyday problems and then incorporate the solutions into the products they design.

 

Image Recognition Takes Center Stage

Image and object recognition is getting a lot of attention lately, judging from the demonstrations by Intel Labs at the Day 0 briefing. It seems that the next big usage model would depend heavily in image recognition especially when it applies to most situations, be it security, education and everyday applications.

The main challenge in delivering great image recognition experience lies with both the recognition algorithm as well as computing performance. The goal is to enable such a feature available in all of Intel's chips, from the most commonly adopted chip like the Atom to processors like the Core architecture. Intel's research is mostly centered around improving the efficiency of the recognition logic while reducing the power consumption of the chip so that mobile applications would not suffer from poor battery life.

Facial recognition demonstrated on a netbook - enabling greater security on all kinds of computing platforms. This is achieved through the use of Intel architecture and internet cloud to enable quicker face recognition.

Besides facial recognition, Intel has also devoted much research resources into enabling mobile devices to recognize buildings and location. Using GPS to narrow the search and comparing its captured images with readily obtainable reference images from Google, the device is able to give a fairly accurate search, complete with facts of the location pulled from rich data sources like Wikipedia.

Educators have always looked to computers to make learning more interesting to today's children. Using Intel's image recognition technology, they have developed a tutoring software for elementary students where a ClassmatePC used its built-in webcam to interact with the student. In this demo, the program dishes out mathematical problems which requires the student to solve them by putting out the correct amount in coins. The program then processes the student's action through its image recognition algorithm.