Ms. Hashani Atapattu, Computer Science (Special) student, won the best research paper presentation award at the 2nd Asia-Pacific Multidisciplinary Research Conference held in Colombo on 29.07.2017.

Her work, based on the final year research project,  presents a computer vision based approach to automatically detect the type of a currency note using a portable device that can be used by a blind person. The proposed approach uses a monochromatic light source to illuminate the currency note in order to segment the currency from the image. A fusion of Bag of Feature of Speeded Up Robust Feature (BoF-SURF) and hue color histogram is used to describe the currency notes and classifying with a Support Vector Machine (SVM). The portable device is setup using a Raspberry Pi 2 model B with a Raspi camera. Based on the type of detection, a feedback is given to the user by specific vibration pattern corresponding to particular currency notes.  The results indicate that the proposed technique segments the currency notes with an average completeness of 95% having a spatial error of 0.3 while keeping the average recognition rate at 91 %.

The project was supervised by Prof. R.G.N. Meegama of the Department of Computer Science, Faculty of Applied Sciences.