A deep learning facial expression recognition based scoring system for restaurants
Download as DOCX, PDF1 like1,104 views
This document presents a facial expression recognition-based rating system for unmanned restaurants, aiming to capture customer feedback through analysis of facial expressions. It highlights the limitations of existing rating systems and proposes a novel approach that encourages all customers to provide ratings using a simple and interactive interface. The system utilizes pre-trained convolutional neural networks (CNN) for face detection and emotion classification, improving the understanding of customer experiences in an automated dining environment.
A deep learning facial expression recognition based scoring system for restaurants
- 1. Abstract: Recently, the popularity of automated and unmanned restaurants has increased. Due to the absence of staff, there is no direct perception of the customers' impressions in order to find out what their experiences with the restaurant concept are like. For this purpose, this paper presents a rating system based on facial expression recognition with pre-trained convolutional neural network (CNN) models. For interactive human and computer interface (HCI) it is important that the computer understand facial expressions of human. With HCI the gap between computers and humans will reduce. The computers can interact in more appropriate way with humans by judging their expressions. There are various techniques for facial expression recognition which focuses on getting good results of human expressions and then the food is supposed to be rated. Currently, three expressions (satisfied, neutral and disappointed) are provided by the scoring system. Algorithms: Haar-AdaBoost Convolutional Neural Network (CNN) Existing System: As there is no staff available in unmanned restaurants, it is difficult for the restaurant management to estimate how the concept and the food is experienced by the customers. Existing rating systems, such as Google and TripAdvisor, only partially solve this problem, as they only cover a part of the customer’s opinions. These rating systems are only used by a subset of the customers who rate the restaurant on independent rating platforms on their own initiative. This applies mainly to customers who experience their visit as very positive or negative. Proposed System: In order to solve the above problem, all customers must be motivated to give a rating. This paper introduces an approach for a restaurant rating system that asks every customer for a rating after their visit to increase the number of ratings as much as possible. This system can be used unmanned restaurants; the scoring system is based on facial expression detection using pre-
- 2. trained convolutional neural network (CNN) models. It allows the customer to rate the food by taking or capturing a picture of his face that reflects the corresponding feelings. Compared to text-based rating system, there is much less information and no individual experience reports collected. However, this simple fast and playful rating system should give a wider range of opinions about the experiences of the customers with the restaurant concept. System Architecture:
- 3. Modules: Face Detection: Face detection or localization is an important step for image classification since only the principal component of face such as nose, eyes, mouth are needed for classification. Face detection algorithms can be broadly classified into feature, knowledge, template and appearance based methods. Our proposed system uses Viola Jones object detection algorithm for face localization which comes under feature based classification. Viola Jones object detection algorithm uses Haar featurebased cascade classifiers. The Haar Cascade classifier is extremely important element of the face detection. The presence of the features in any of the input image is determined by the Haar features. Facial Expression Recognition classification: After learning the deep features, the final step of FER (Facial Expression Recognition) is to classify the given face into one of the basic emotion categories. Unlike the traditional methods, where the feature extraction step and the feature classification step are independent, deep networks can perform FER in an end-to-end way. Specifically, a loss layer is added to the end of the network to regulate the back-propagation error; then, the prediction probability of each sample can be directly output by the network. In CNN, softmax loss is the most common used function that minimizes the cross-entropy between the estimated class probabilities and the ground truth distribution. Convolutional neural network (CNN): CNN has been extensively used in diverse computer vision applications, including FER. At the beginning of the 21st century, several studies in the FER literature found that the CNN is robust to face location changes and scale variations and behaves better than the multilayer perceptron (MLP) in the case of previously unseen face pose variations, employed the CNN to address the problems of subject independence as well as translation, rotation, and scale invariance in the recognition of facial expressions.
- 4. SYSTEM CONFIGURATION: Hardware requirements: Processer : Any Update Processer Ram : Min 4 GB Hard Disk : Min 100 GB Software requirements: Operating System : Windows family Technology : Python 3.6 IDE : PyCharm