1. Biometrics
This research cluster currently focuses on using biometrics for the re-identification of both human and animal.
Human re-identification
Gait: Gait is one of the well-recognized biometric features to re-identify human at a distance from a camera, by inspecting their walking manners. It is more suitable for the surveillance applications, when compared with other popular biometric features such as face and fingerprint. This is because these two features must be very close to a camera or need the physical contact to the reader-machine. The practical challenges of the gait recognition are worked out here, including the view changes and walking speed changes.
Face: The face recognition technology is extended here for the practical use of the re-identification of the security guards under the real environment of day and night time.
Animal re-identification
The muzzle print is used as a biometric for the automatic cattle identification.
2. Vision-based Human Interaction
Recently, the user interaction is not limited to the traditional media such as mouse, keyboard, joint stick, or handhold controller. The natural movement is used as the controller via various type of cameras such as webcam and Kinect. For example, the traffic hand signals are recognized using the real movements of the user via Kinect. This is used as the controller for the decision-making training in the traffic simulator.
Another project is about the game-based enhancement for the rehabilitation. The real movement of the patient is used as the controller of the game, based on four fundamental movements including shoulder flexion, shoulder abduction, shoulder horizontal abduction, and elbow flexion/extension. These actions are recognized via Kinect in four designed games.
3. Medical Image Processing
This is for segmenting key/interested components in the medical images. The first sample project is to segment the key signals of the diabetic retinopathy such as microaneurysm and exudate. The second sample project is the segmentation of abdominal aortic aneurysm. We focus on both 2D and 3D domains.
4. Elders/Disabilities Support
This cluster is to develop the applications to support the daily living of elders and/or disabilities. For example, we developed the applications to support the learning for blind, including the English-Thai Braille Dictionary and the Braille Graph Builder. The English-Thai Braille Dictionary is an android application to facility the use of English to Thai dictionary by allowing the user to input English words via a Braille keyboard on the screen.
The Braille Graph Builder is developed to support both visually impaired students and instructors to have a better experience in teaching and learning graphs and equations. The text-to-speech is employed to speak out all working processes during the program usage, in order to facilitate the visually impaired users.
The next project is the development of the game prototype for the rehabilitation enhancement by using computer vision technique to detect, track and recognize the actions of the player in the free-controller and free-tracker environment. All actions are based on the selected activities from the activities of daily living (ADL). The gameplay is recorded for the further evaluation and analysis by the therapist.
Also, the game-based for enhancing communication skill of children with autism spectrum disorder is developed and evaluated on the real cases.
In addition, the Eldercise program is developed as an android application. It is designed to help encouraging elders/ middle-aged office workers to exercise properly in an easy way using the smart phone. The accelerator in the smart phone is trained to recognize the right actions of six exercises including arm raise, triceps extension, swing arms, stretch arms, shoulder tap, and biceps curl.
5. Games/Mobile/Web Application
This cluster covers the games and applications that are not designed specifically for elders or disabilities. For example, the turn-based strategy board game is developed mainly for the entertainment purpose. Various movement AI techniques are implemented for the different levels of the game.
I’m Road, Fury Traffic is the car running game for the entertainment purpose which also evokes conscious mind of good practices in driving car.
Also, the Interactive Storytelling Game (ISG) and the Digital Game-based Learning (DGL) are developed and evaluated to prove the concept of using game as the learning media to delivering the technical knowledge to general people. The simulator for the medial field is also developed for the purpose of increasing the awareness and experience in the orthopedic ER.
6. AI/CV based Application
This cluster focuses on using AI and/or CV in solving real-world problems, including the optimization for location routing problem in supply chain management, the water quality prediction in shrimp culture, the co-citation analysis for identifying trends of interdisciplinary research, and the 3D bone reconstruction using Kinect.
