Mobile Application of Text Recognition

Abstract

In the project, an IOS app is developed for the purpose of text recognition. Through this app, the user can easily recognise the text on any printed materials just by capturing a photo with their phone camera. The recognised result can be saved, exported as text files and further shared to other online platforms. The highlights of the project are the applications of various image processing techniques. For instance, filters, convolution calculations, averaging and thresholding of horizontal histograms. By applying these methods, background and non-text objects of the image are filtered and noises of the histograms are largely eliminated to facilitate the targeting of the text. Through these processes, an option is provided for the user to select whether recognise the whole image or just parts of it in order to enhance the flexibility of the recognition, and also improve its accuracy.