| Abstrakt: |
Presently, a majority of businesses are carried out online, and there is a trend of creating a virtual environment during online shopping. The recent advancements in the fields of e-commerce and business have necessitated providing personalized online recommendation to the users or customers by predicting human emotion. Though it is easy to predict human emotion by observing one's facial expression, but achieving the same with computer algorithm is challenging. In this work, a novel approach has been proposed for online recommendation system using human facial expression-based emotion detection. The facial expression is taken in real time, and thus, the proposed framework provides an intelligent recommendation system on–the-fly without relying on historical ratings or purchase records. This new approach is to predict the human emotion based on the facial features and to develop ways for recommending products to customers based on such reaction during the purchasing of a product. Two approaches have been used in this work. First, with the help of five different facial expressions, emotion of the user is detected, and analysis is done. Secondly, video of user's facial expression is captured through a webcam while purchasing a product online. To predict emotion in real time, the live video from a webcam is fed into a network which detects faces and predicts the emotions. This proposed system can detect the emotion with 75% accuracy and provide online recommendation in real time. The performance measurement of the proposed recommendation system is done by taking feedback from the customers. The purpose of this work is to create a virtual environment and to provide recommendations in real time using deep learning approach. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink