Project Description

EEG Based Media Rating System

Alfadhl Abbas Ali
Computer Engineering Student
Sayed Ahmed Abbas Ahmed
Computer Engineering Student
Sayed Murtadha Adnan Redha
Computer Engineering Student
Supervised By:
Dr. Ali Hasan
 Assistant Professor
Abstract

NeuroRate is a cross platform mobile application that incorporate several cutting edge tools and technologies such as Flutter framework, NeuroSky MindWave Mobile electroencephalogram ( headset, Microsoft Azure Machine Learning Studio, Amazon Web Services, Unity game engine, and Samsung Gear VR headset The application offers a novel platform for rating media content by analyzing human emotions while watching media such as movies, games, educational videos, etc The user of NeuroRate would require an EEG headset to feed real time emotion signals to the cloud where an AI engine is used to classify the emotional state of users while watching the media This provides a new method for media rating and analysis Moreover, it reduces some major critical points on current rating systems including system reliability and integrity of the reviewers The application serves two kinds of users publishers, and reviewers The publishers can take advantage of the application by uploading media content and requesting/inviting several registered reviewers for rating The registered reviewers have position of special EEG headset Reviewers who accept rating requests can view and perform rating roles while utilizing this EEG headset Currently, the system has been prototyped for analyzing four emotional states which are sadness, happiness, excitement, and natural emotions

Objectives
  1. To propose a system that consists of the brain to computer interface device that will be held on the reviewer to analyse his/her emotion while he/she watching a media content.
  2. To propose a novel automated rating system that depends on the emotional state of the reviewers.
  3. To implement the proposed system in cloud.
  4. To develop an AI model and deploy it in cloud to be used in the rating system.
  5. To serve two types of users: publishers, and reviewers.
  6. To utilize the newest trends in technology while developing the system.
Methods/Diagrams/Figures
Results
  1. The system enhances the media rating procedure experience.
  2. The system provides average accuracy of 75 in theclassification process for the current provided datasett.
  3. The system supports multiple types of media contents such as movies, educational videos, and TV shows, etc.
Conclusion and Future Work
  1. Support more classes of emotions to match different categories of available media contents.
  2. Expand the scope of the system to support music, games, and books rating.
  3. Improve overall system UI depending on users’ feedback.
  4. Fully migrate the system to AWS.
View Poster Online