PhotoMath: Math Made Too Easy?

Oct 27, 2014 By Deepa Gopal
Deepa Gopal's picture

The thought of Math can be distressing for some, with complex equations sending them into a tizzy! Now, imagine an app that can automatically solve those nasty math equations for you.

PhotoMath, announced recently in London, has taken the app world by storm. It has become the number one free app in the App Store for most downloads.

How does it work? Simply move your smartphone's camera over a mathematical expression or an equation asking you to solve for 'x'. Take a photo and presto! You have the answer. Not just that, you also get to see the steps it takes to arrive at the solution. For now, the app only works on printed text, and not handwritten equations. 

The Technology

PhotoMath uses a technique called Optical Character Recognition (OCR) to 'parse' the mathematical expression, and solve it appropriately. OCR converts images of typewritten or printed text into "machine-encoded text" - a fancy term for a .txt or .doc file.

OCR works in two ways: Pattern matching and Feature extraction. Pattern matching involves comparing an image to a known character pixel-by-pixel. By comparing the pixels to those of known characters, it is possible to classify the image as one of those characters (say the letter 'd'). This approach is simple and works best with typed text. But it will not perform well for different fonts or handwritten characters.

Feature extraction is a much more sophisticated approach. Here, the character is analyzed and classified based on important features like lines, curves, loops, etc. Then, this 'abstract-ified' text is compared with a similar representation of a known character. This approach, although harder to implement, is much more sophisticated and is used in most modern OCR technology.

PhotoMath relies on these techniques to convert the image into something the computer can understand, and then spits out the answer to the equation impressively fast.

A Teacher's Worst Nightmare?

There is concern from parents and teachers that the app could be used by students for dishonest purposes. MicroBlink, the company behind this technology, insists it is an educational tool. According to Izet Ždralovic, the co-founder of the company, "If someone is a cheater they can always find a way to cheat. They can put the (question) on the Internet. Cheaters can use it to cheat but if you see the potential you can use it to learn.".

While PhotoMath is currently able to solve problems through a 9th or 10th grade level, the company is working to improve it and also add higher levels of math. There are bugs with the current software that need fixing. The company also plans to use the technology in other fields, and to find word definitions, or help read handwriting.  

What do you think? Will PhotoMath encourage learning, or cheating?

Courtesy Mashable, explainthatstuff.com