Classification technique for human emotion in virtual reality using game-based brain computer interface

• Main Author: Abuhashish, Faris Amin Muflih
• Format: Thesis
• Language: English
• Published: 2015
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://eprints.utm.my/id/eprint/54837/1/FarisAminMuflihPFC2015.pdf; http://eprints.utm.my/id/eprint/54837/; http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:88011

The substantial amount of reviews in the realms of computer graphics and the multimedia as well as emotion synchronizing and controlling techniques of 3- Dimension (3D) have thrown the 3D Virtual Human (VH) model in Virtual Reality (VR) into the spotlight. It only requires a small number of 3D VH model systems to manage emotions through sophisticated procedures that include human brain activity together with 3D emotion expression feedback. However, this circumstance leads to a deficiency in emotion interpretation. Emotion interpretation is crucial for the categorization of human sentiments so that they can be coordinated and plotted with a 3D VH model to generate the interaction outcome via emotional walking expression and reveal complete emotion interaction feedback in VR. This study recommends a hybrid emotion classification technique which attains the immersion of emotion interaction with a 3D VH model. This technique involves three steps. Firstly, the criterion of the obstacle that requires a solution is identified. The second step involves emotional feature extraction through a reformulated method, and categorization with a hybrid method and plotting with a defined formula. The third step entails the assimilation and execution of all the features of the recommended technique and mapping the classified emotions. This includes the implementation of a synthesis of emotional walking alongside emotion integration, brain activity and the 3D VH model. Ultimately, the recommended model is analysed and substantiated through actual emotion effects on the 3D VH model with emotional walking style in a VR circumstance. The classified accuracy percent is 88.7% that is achieved by the proposed technique. Outcomes from the tests established that the enhancement of immersion of emotional expression through this procedure is achievable through the utilization of game-based Brain Computer Interface (BCI) in the VR domain. The employment of this technique considerably elevates the realism and immersion of other applications such as robotics regarding emotion.