International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 11 | Nov 2024
p-ISSN: 2395-0072
www.irjet.net
A SURVEY ON FACE EMOTION RECOGNITION AND ANALYSIS Robin Nadar M.Tech. Student, Electronics and Telecommunication, Cognitive Institute of Interdisciplinary Studies, Mumbai, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The objective of the Face Emotion Recognition
specifically the VGG-Net, when paired with the FER2013 dataset. Their work demonstrated the ability to achieve emotion recognition in real time, which is a crucial feature for many applications[1].
and Analysis system is to identify people's emotions in real time and record them in a file. This information can then be examined and analyzed to learn more about the individual whose emotion is being identified. A timestamp is used to record the feelings, which aids in classifying the data based on the various periods of the day. This information can be useful in observing patterns in emotions throughout the day to get insight into an individual's emotional state. As a result, the objective of this paper is to provide an accurate analysis of the numerous Face Emotion Recognition and Analysis systems currently on the market, as well as the datasets that may be utilized for training and testing. A list of methods was gathered and algorithms that can and are often used for Face Emotion Recognition and Analysis, which was examined. A comparative study we also conducted of the assessed systems' accuracy and efficiency, as well as their different limitations.
Similarly, Azizan and his team investigated various classification methods, including Haar Cascade and Softmax, to see how they can be applied to tasks related to categorizing emotions. Their findings contribute to understanding which techniques work best for identifying different emotional states[2]. Another important study by Dores and his colleagues examined how gender differences might influence the effectiveness of emotion recognition systems. They highlighted the necessity of using diverse datasets to ensure that these systems can provide accurate results across different populations. This finding emphasizes the importance of inclusivity in the training data used for these systems[3].
Key Words: Face, Emotion, Recognition, Analysis, CNN, Convolutional Neural Network, Neural Network, FER, Face Emotion Recognition and Analysis.
Furthermore, Dachapally discussed the benefits of using autoencoders and deeper convolutional neural networks (CNNs) to improve the accuracy of emotion detection. This indicates a trend toward more complex models that can achieve better outcomes[4].
1.INTRODUCTION Facial emotion recognition is a growing area of research that brings together three important fields: artificial intelligence, psychology, and human-computer interaction. This area focuses on how emotions can be understood by analyzing facial expressions. By interpreting these expressions, researchers can gain valuable insights into human behavior, which can be applied in several important areas such as monitoring mental health, improving customer experiences, and enhancing security systems.
Ali and his research group provided an in-depth analysis of how CNNs utilize specific activation functions, like ReLU, and optimization techniques, such as Adam. These methods significantly enhance the performance of models when it comes to tasks involving emotion detection[5]. Overall, these references form a strong foundation for this research survey. They offer a systematic way to understand the current state of facial emotion recognition systems and provide insights on how to improve them in the future.
In recent years, the development of deep learning techniques and improvements in computer vision have greatly increased the capabilities of facial emotion recognition systems. These advancements make it possible for machines to recognize and interpret emotions more accurately and efficiently than before. This research aims to thoroughly examine the basic elements required to build effective facial emotion recognition systems. This includes studying the datasets that are available, exploring existing algorithms designed for emotion recognition, and evaluating how well these systems perform in real-life situations.
2. DEEP LEARNING [1] Deep learning is a machine learning technique that gradually extracts higher-level properties from raw input. whereas the word deep refers to the utilization of multiple layers in the network. A linear perceptron cannot be used as a universal classifier, but a network with a nonpolynomial activation function and one hidden layer with infinite width may. Deep learning is a modern form that focuses on layers of predetermined size, allowing for practical application and fast implementation while maintaining theoretical
A comprehensive review of prior studies in this area shows significant advancements made over time. Sinha and colleagues illustrated the success of deep learning models,
© 2024, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 731