Optical Character Recognition is the conversion of scanned or photographic images into editable, machine-readable text that can be searched.
OCR works by submitting an image to an OCR engine. The engine works by matching pieces of the image provided to shapes it is instructed to recognize. For instance, if you were to make the letter "V" with your index and middle finger (also known as a 'peace sign') and submitted that photo to an OCR engine, the engine would identify that image as the letter "V", because the shape of your fingers best matches the letter "V", (assuming the engine is using the English alphabet.) The engine would then return the letter "V" as editable, plain text.
When filling out forms or entering secure areas online, you may have noticed those annoying confirmation boxes that force you to read something from an image and enter in the text as confirmation that you're a human being and not some type of web-bot. The reason these are in place is because of brute-force hacking. A brute-force hack, or multiple hack attempt would be someone setting up an automated macro process that repeatedly tries to log in to an area with a username thousands of times. It would have a dictionary with thousands of common passwords, and it would not stop until the list was depleted. Another reason is because there are automated "spiders" (or bots) patrolling the net for web pages to index. The confirmation box you see is called CAPTCHA, and it ensures that you are not a bot. By being able to read the text from the image with your eyes, and manually typing it into the box, you are essentially acting as a human OCR engine. You are converting an image into machine-readable text.
The annoying thing about them is that they are hard to read sometimes, and sometimes it takes multiple attempts to crack what should be an obvious entry code. They throw in symbols and break up the text to confuse bots that are OCR-capable, it's purpose is to confuse the engine and render the bot useless, forcing it to move onto another area.
Sources:
1. An Introduction to Optical Character Recognition
2. CAPTCHA
OCR works by submitting an image to an OCR engine. The engine works by matching pieces of the image provided to shapes it is instructed to recognize. For instance, if you were to make the letter "V" with your index and middle finger (also known as a 'peace sign') and submitted that photo to an OCR engine, the engine would identify that image as the letter "V", because the shape of your fingers best matches the letter "V", (assuming the engine is using the English alphabet.) The engine would then return the letter "V" as editable, plain text.
License plates can also be read using OCR technology
One major use of OCR technology can be found in any text-to-speech application, which are used by people with visual disabilities. They aren't able to read text from the screen, so the computer reads it out loud for them. The OCR engine goes through any given web page or document, converts the words into text, and runs it through another program as input, and then outputs the text as audible speech.
Some companies have also mastered the concept of OCR check deposits, allowing bank users to deposit checks into their accounts simply by taking a photo of the check with a cell phone, and uploading that image to the bank's server. As with the text-to-speech, there are two steps involved here, first is submitting the image of the check as input to the OCR engine, and then handing off the output of the engine as input into the bank's server where it can run the text against their internal database and handle the data like it normally would.When filling out forms or entering secure areas online, you may have noticed those annoying confirmation boxes that force you to read something from an image and enter in the text as confirmation that you're a human being and not some type of web-bot. The reason these are in place is because of brute-force hacking. A brute-force hack, or multiple hack attempt would be someone setting up an automated macro process that repeatedly tries to log in to an area with a username thousands of times. It would have a dictionary with thousands of common passwords, and it would not stop until the list was depleted. Another reason is because there are automated "spiders" (or bots) patrolling the net for web pages to index. The confirmation box you see is called CAPTCHA, and it ensures that you are not a bot. By being able to read the text from the image with your eyes, and manually typing it into the box, you are essentially acting as a human OCR engine. You are converting an image into machine-readable text.
An example of a CAPTCHA
The annoying thing about them is that they are hard to read sometimes, and sometimes it takes multiple attempts to crack what should be an obvious entry code. They throw in symbols and break up the text to confuse bots that are OCR-capable, it's purpose is to confuse the engine and render the bot useless, forcing it to move onto another area.
Sources:
1. An Introduction to Optical Character Recognition
2. CAPTCHA
No comments:
Post a Comment