Group Analysis in Real Time, Detecting the Emotion of a Person in a Meeting Using Image Processing

Real-time data is highly valued in Crowd Analysis, which involves the collection, aggregation, and application of information related to people's daily lives. However, many current applications do not operate in real-time. Our aim is to develop a unique image processing application that is fully automated, cost-effective, and capable of achieving human- level or better performance without human intervention, a capability that most existing applications lack. To this end, we propose a web application that uses a neural network to automate the group discussion process by analyzing real-time camera feeds and detecting the emotions and attentiveness of participants. This analysis involves three steps: pre- processing, object detection, and event/behavior recognition, and determines a person's nature through the analysis of their emotions. The application can help identify outstanding candidates in group discussions and make the recruitment process more efficient. Keywords associated with this project include emotion recognition, face expression recognition, feature extraction, patched geodesic texture transform, neural network, curvelet feature extraction, bag of words method, local directional number pattern, regional registration technique, and gradient feature matching.