Teaching Machines to Read Characters and Symbols
People have many different ways of communicating with others, such as facial expressions, tone of voice or behavior. Within these various communication methods, characters and marks can only be used by humans.In our daily lives, we absorb all types of information through characters and marks written on paper or signboards. However, with the rapid advancement of information technology, it has now become essential to access and handle information in electronic format with computers and other information equipment.
Increasing demand for simpler information input.
When using computers and other equipment to process information, the biggest bottleneck is at the input stage, in which input is still performed manually in most cases. This process of entering huge volumes of information by hitting keys and flipping pages is extremely tedious and time-consuming. As PCs have become essential tools at offices and schools, the need to resolve this impasse is becoming even more critical. One solution is the automatic reading of information from a scanned image. Optical Character Recognition (OCR) and mark recognition technologies are essential for making this possible. OMRON has committed substantial R&D resources to the creation of systems that can recognize marks and characters in an image. The result is what OMRON calls "Pattern Vision."
Information input is faster and easier than ever.
One example of a product developed using Pattern Vision is "OmCR," a sophisticated type of character recognition software for PCs. The software can detect English and Japanese characters from scanned various documents and convert them into electronic format. OmCR is recognized as an industry leader in performance among other character recognition software on the market. It is also embedded in a popular machine translation software application for character recognition.
"OmCR Ver.3" software from OMRON Software Co., Ltd.
OMRON's Pattern Vision technology has also led to the development of a highly miniaturized character recognition unit on a chip. This is expected to find numerous applications, an example of which is cellular phones. In Japan, cellular phones with built-in cameras have already become essential commodities for average consumers. What kind of convenience can a cellular phone equipped with this tiny chip offer to users?
Let's take an example of advertisements we see in trains or on the street. Nowadays, most of these advertisements have a URL for the company's website. TV programs also often show URLs and other superimposed text such as addresses and telephone numbers. But how many people actually bother picking up a paper and pencil to write them down? While you're looking for something to write with, the text may already be gone. Wouldn't it be handy if you could take a picture of the information with your cellular phone, which could then automatically convert the information in the image into text data? This type of intelligent cellular phone will become a reality in the not-too-distant future.
Business cards are essential for all types of workers. But everyone knows how troublesome it is to have to input new data into a telephone or e-mail directory every time you get a new card. Using a cellular phone with character recognition, you can simply take a picture of the card and convert the information into electronic format. This is a huge time-saver, and also reduces the chance of inputting information incorrectly.
Character recognition from a cellular phone camera image