The following two research projects of Faculty of Computing, KDU are in the “NBQSA 2020 Winners’ Circle (https://www.nbqsa.org/winners-circle-nbqsa-2020/)” under the tertiary category.

1. An Automated Robotic System for Traditional Drums in Sri Lanka by Ms.M.A.S.T Goonatilleke (Supervisor – Dr. B Hettige)
   Sri Lanka has a precious music culture that is based on traditional drums. However, at present, this tradition is gradually dying out due to a lack of talented drum players. Therefore, can propose a better solution using new technologies to solve this problem. This project is about an automated robotic system that was designed for the Thammattama which is one of the major four drums in Sri Lanka. This system consists of two robotic arms that are carrying two sticks of the Thammattama and play standard drum tunes correctly and efficiently like a drum player without any intervention of a human. This system has been designed using robotic technology and can be replaced instead of a drum player to play standard drum tunes that are playing using the Thammattama. The system is very user-friendly as it operated using only two switches as ON/OFF the system and change drum tunes. In addition, any user in  anywhere can operate this system easily and quickly via a mobile app.

2. Mobile App for Automatic Identification of Snake Types Found in Sri Lanka Using Convolutional Neural Networks by Savini Abayaratne, (Supervisor- Ms. WMKS Ilmini (FOC, KDU) and Dr. TGI Fernando, Senior Lecturer, University of Sri Jayewardenepura)
   The anti-venom that should be given to a snakebite victim differs from a snake to snake and therefore identifying the snake is essential in treating the patient. This mobile application was developed to address this problem where it automatically identifies six snake types (Cobra, Common Krait, Saw Scaled Viper, Russell’s Viper, Hump-nosed Pit Viper, and Rat Snake) found in Sri Lanka. The mobile application employs a Convolutional Neural Network (CNN) model trained using transfer learning and fine-tuning with the MobileNet pre-trained network and a dataset of 12,000 images. The mobile application is capable of classifying an image in real-time, a static image captured from the camera while using the application, and an image stored in the gallery. The classified result is shown in both Sinhala and English along with the accuracy rate. Once a user identifies a snake, he/she has the option to learn more about that particular snake through the “about this snake” option provided in the application. The mobile app also has a feature to directly place a call to contact an authority listed in the application and other options to get more information on snakebite prevention, first aid that can be administered to a victim, and information on the six snake types.