AI Mock Interview Behavioural Recognition Analyst
Dr.Reshmi B
Dept Of Cse
Ahalia School Of Engineering And Technology
Palakkad
reshmi.b@ahalia.ac.in
Alent Eldho
Dept Of CSE
Ahalia School Of Engineering And Technology
Palakkad
alenteldho123@gmail.com
Dr.Gunasekaran Subramanian
Dept Of CSE
ahalia school of engineering and technology
palakkad
Glorin Jose
Dept Of CSE
Ahalia School Of Engineering And Technology
Palakkad
glorinjose605@gmail.com
Rohit S
Dept Of CSE
Ahalia School Of Engineering And Technology
Palakkad
rohits@gmail.com
Abstract : This proposes a comprehensive framework that integrates a emotion detection , facial recognition, and AI-driven mock interview analysis based on the deep learning techniques. The system utilizes Semantic-Emotion Neural Networks (SENN) to understand emotions from text-based data, YOLO and ArcFace for accurate facial recognition, and a combination of Convolutional Neural Networks (CNN) and Natural Language Processing (NLP) to analyze interview answers. Through the incorporation of these advanced models, the presented system attains improved accuracy and real-time performance in various applications.
To measure its efficiency, the system was subjected to a benchmark data tests, displaying more accurate and better performance in relation to state-of-the-art algorithms. The paper also presents an in-depth look at multimodal deep learning methodology in augmenting human behaviour’s recognition and measurement in the online space.
Deep learning and artificial intelligence have transformed the field of human-computer interaction, particularly emotion recognition and behavioral measurement. Contemporary neural network models support real-time emotion detection from audio, text, and visuals. Face recognition has also evolved considerably with deep metric learning, which improves classification accuracy and enhances feature representation. By integrating these technologies, the suggested multimodal framework efficiently addresses essential challenges in real-time human behavior analysis, achieving the best possible trade-off between performance and computational efficiency