Automatic Personality Recognition In Interviews Using CNN

Due to the pandemic, various industries are facing significant challenges. The IT sector is also struggling with recruitment processes due to the difficulties associated with analyzing candidates online. Consequently, automated video interview analysis has emerged as an active research topic, which can help identify specific personality traits. Convolutional neural network models based on DL techniques have been developed due to advancements of CV and pattern recognition, and this technology has various applications in areas such as personality computing, psychological testing, and human computer interaction. These models accurately detect nonverbal cues and ascribe personality traits to individuals using a camera. With the help of AI-based interview agents, organizations can either replace or supplement the current self-reported personality evaluation tools that job candidates often manipulate to obtain socially acceptable results. The development of an AI-based interviewing system involved the use of asynchronous video interviews (AVIs) and the application of a personality prediction model trained on the first impression v2 dataset. The objective was to achieve automated personality recognition (APR) through the extraction of relevant features from the AVIs and the use of genuine personality scores obtained from facial expressions. This was achieved through the utilization of the VGG-16 network to train the model for improved accuracy in personality prediction.