JavaScript Dhivehi Character Recognition

Here is another of my pet projects brought back from the land of the deceased.

This one is called "JavaScript Dhivehi Character Recognition". It was created early 2003 (or maybe late 2002) and made available on Basically, it lets you draw a Thaana character using your mouse and then it "recognizes" what you have drawn. The purpose was mostly to satisfy my curiosity into artificial intelligence and pattern recognition at the time, however it also showed promises of the beginnings of a future where Dhivehi documents maybe scanned in and processed by a computer to convert it to text just as Optical Character Recognition technology has been doing for English documents. I think this rudimentary application was the first ever Dhivehi character recognition implementation released to the public. More interestingly, this seems to be the only character recognition implementation programmed in JavaScript floating around on the Internet even now. :-D

I spent a bit of time tonight reworking some bits of the code for clarity. The entire implementation is done using JavaScript and DHTML. You are welcome to study the code to see how it works. The code is well commented and maybe a good starter into AI and pattern recognition basics. It uses a single layer single Perceptron model to really simplify things however it is a good enough practical implementation to work for characters drawn on a 10x10 grid. The grid makes up the input data to the neural network. The neural network is hard-coded into the page and has definitions for each character in the alphabet. I do hope you are surprised by the accurateness of the recognition of this little application.

Have a look at it HERE. Let me know if you find it amusing... or not.

My company - Technova Pvt Ltd - is currently working on bringing a full fledged Dhivehi OCR software to the Maldivian public. It will probably be made available early 2006, as a service for customers requiring bulk OCR processing. We shall be releasing Windows, Linux and Mac versions of the software for home and business use around mid 2006.