International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 04 | Apr 2024
p-ISSN: 2395-0072
www.irjet.net
Face Emotion Recognition System Using Deep Learning ASST. PROF M.KAVITHA1, V. HARSHAN2, M. PRAVEEN KUMAR3, A. RAHUL4 1234 Dept. of Computer Science and Engineering, Government College of Engineering Srirangam, Tamilnadu, India
---------------------------------------------------------------------***--------------------------------------------------------------------1.1 RELATED WORK
Abstract - This paper presents an enhanced system for real-
time facial emotion detection, aiming to improve efficiency and accuracy through deep learning. The proposed approach utilizes VGG-19 transfer learning, a pre-trained convolutional neural network (CNN) architecture known for its depth and strong performance in image classification. VGG-19's pretrained weights contribute to improved efficiency compared to simpler CNNs, allowing for effective feature extraction and classification of emotional expressions in real-time. This approach has the potential to benefit various applications in human-computer interaction and psychology by enabling accurate and timely emotion recognition
Sharmeen M. Saleem Abdullah and Adnan Mohsin Abdulazeez [13] addressed the latest FER analysis. Numerous CNN architectures have been identified that have recently been proposed. They have provided various databases of random photographs obtained from the actual world and other laboratories to detect human emotions. Hussein, E. S., Qidwai, U. and Al-Meer, M. [4] recommended a CNN model to understand face emotions with three continuum emotions. This model uses residual blocks and depth-separable convolutions inspired by Xception to minimize the sum of parameters to 33k. They use a convolutional neural FER network for emotional stability identification. CNN uses convolution operations to learn extract features from the input images, which reduces the need to extract features from images manually. The proposed model offers 81 percent total precision for invisible results. It senses negative and positive emotions, respectively, with a precision of 87% and 85%. However, the accuracy of neutral emotion detection is just 51%.
Key Words: Facial Emotion Detection, Deep Learning, VGG19 Transfer Learning, Real-Time Emotion Recognition, Efficiency, Human-Computer Interaction, Psychology
1.INTRODUCTION Human communication encompasses speech, gestures, and emotions, vital for interpersonal interactions. AI systems capable of understanding human emotions are crucial, especially in healthcare, and e-learning where emotional understanding is paramount. Traditional emotion detection methods often fall short in real-time scenarios, necessitating models that can continuously interpret facial expressions for dynamic emotional assessment.
Jiang, P., Liu, G., Wang, Q., and Wu, J [5] introduced a new loss feature called the advanced softmax loss to eradicate imbalanced training expressions. The proposed losses guarantee that any class would have a level playing field and potential using fixed (unlearnable) weight parameters of the same size and equally allocated in angular space. The research shows that proposed (FER) methods are better than specific state-of-the-art FER methods. The proposed loss can be used as an isolated signal or used simultaneously with other loss functions. To sum up, detailed studies on FER2013 and the real-world practical face (RAF) databases have shown that ASL is considerably more precise and effective than many state-of-the-art approaches..
This paper proposes a real-time facial emotion recognition model leveraging AI and computer vision advancements. The model aims to enhance human-computer interactions across diverse applications by dynamically detecting and responding to emotions. Automatic Facial Expression Recognition (FER) has gained traction, driven by its potential in human-computer interaction and healthcare. While Ekman's discrete categorization model is widely used, its limitation in handling spontaneous expressions prompts the need for more comprehensive approaches.
2 METHODOLOGIES This approach utilizes the VGG19 architecture for Facial Emotion Recognition by preprocessing the dataset, training the model, and validating it for real-time deployment. It includes implementing a user interface for interaction and feedback loops for continual improvement.
Our focus is on categorical facial expression classification using the VGG-19 model, known for its depth and performance in image tasks. By employing pre-trained weights, our system achieves efficiency and accuracy for real-time emotion recognition. This work explores VGG-19 transfer learning's potential for facial emotion recognition while remaining adaptable to other models with suitable data.
© 2024, IRJET
|
Impact Factor value: 8.226
CNNs, or Convolutional Neural Networks, are crucial in deep learning and particularly effective in computer vision tasks. They automatically learn relevant features from raw input data, making them ideal for image and video recognition. Structured to mimic human visual processing,
|
ISO 9001:2008 Certified Journal
|
Page 2019