LLEETS Computer Vision+Outsourcing

Product overview: LLeets is a state-of-the-art text recognition application. This multilingual computer vision app lets you detect and extract various types of text, both digital and handwritten. The customer wanted to make this app so easy and fun to use that people would prefer to take a picture of something and use the caputured text instead of writing something down.

Issues: The customer wanted to outsourcing the development of the entire project. Five of our professionals were involved in its creation. We needed to create the app, which allowed users to detect and extract data from any type of document, ranging from digital surveys to basic handwritten notes and letters. The app must be able to work with different surfaces and backgrounds, for instance, digital photos, analog photos, white paper, or yellow sticky notes. The app should be multilingual, covering seven languages: English, German, French, Italian, Russian, Chinese, and Norwegian. We had to incorporate both formal and informal language. We needed to gather big data from all currently available sources and find suitable algorithms there were pre-built or develop them from scratch, and implement special computer vision algorithms to accurately recognize written data.

Technologies used: computer vision

Result: LLeets is robust, easy to use and effectively performs all its functions.