Image enhancement for text extraction on mobile application

• Main Author: Fiona, Kandou.
• Format: Final Year Project Report
• Language: English; English
• Published: Universiti Malaysia Sarawak, (UNIMAS) 2015
• Subjects: QA76 Computer software
• Online Access: http://ir.unimas.my/id/eprint/39562/1/FIONA%20%2824%20pgs%29.pdf; http://ir.unimas.my/id/eprint/39562/4/Fiona%20Kandou%20ft.pdf; http://ir.unimas.my/id/eprint/39562/

Images are captured using a mobile phone camera would create 2D image variants, such as uneven lighting, cluttered background and text-like object. Pre-processing of images aim to reduce the noise in the image. In this project, a mobile application on image enhancement is developed. The application uses a combination of Otsu's Threshold binarization and colour clustering method is proposed to separate the text from the background. Text in images often has a higher intensity or lower intensity than the background. Otsu's binarization is used to convert the image to binary image for a higher contrast between the text and background. However, Otsu's binarization cannot distinguish between different colour pixels that have the similar illumination. Colour clustering is used in conjunction to binarization method as it is able to differentiate the different colour pixels hence eliminating the problem of Otsu's binarization method.