LCD/LED Digit recognition by iPhone

Home medical devices with LCD or LED screens are quite common. While they offer many conveniences, the data is usually not retained. In this thesis, an iPhone application will be introduced that recognizes the digits on LCD or LED screens. The user takes a picture of the desired LCD or LED, and the app will convert the image to text in seconds. The app also has an email feature that allows the user to send the picture and the text conveniently to a secure medical data logging database. A contour finding algorithm is illustrated in this application for image preprocessing. This algorithm is relatively faster and more efficient than typical image preprocessing techniques. The Tesseract optical character recognition engine will be implemented for recognition of digits. A two-step recognition process improves the accuracy of the recognition. In addition to the image processing techniques, cross-compiling, Objective-C coding and Cocoa TouchTM event handling are discussed in this application.