The detection of emotional states from human facial expressions is a crucial research area in artificial intelligence and computer vision.This paper explores real-time emotion recognition using the EfficientNet-B0 model on the FaceEmocDS dataset.The FaceEmocDS dataset comprises 72,412 high-quality images, encompassing eight emotion classes: anger, contempt, disgust, fear, happiness, neutral, sadness, and surprise.Derived from multiple open-source datasets, it ensures diversity and balance after rigorous cleaning to remove lowquality or redundant samples.The model was trained via transfer learning, leveraging pre-trained weights from ImageNet, and fine-tuned over 30 epochs on the PyTorch framework.It achieved 74.32% accuracy on the validation set.To enhance generalization, data augmentation techniques were applied, including random rotations (up to 10 degrees), color jittering (adjusting brightness, contrast, and saturation), random resized cropping, horizontal flipping, and random erasing (with probability 0.7).Class weights were incorporated into the CrossEntropyLoss function to address imbalances, particularly in underrepresented classes like contempt and disgust.For real-time implementation, MediaPipe Face Detection was integrated for efficient face localization, followed by emotion classification with the EfficientNet-B0 model.The system operates at 20-30 frames per second (FPS) on standard hardware, making it suitable for dynamic environments.Test results yielded an accuracy of 73.18% and a weighted F1-score of 0.7327, with highest performance on happiness (93.08%) and lowest on contempt (63.05%).Confusion matrix analysis revealed common misclassifications, such as contempt with anger or disgust, highlighting areas for improvement.Compared to baseline models such as VGG and ResNet, this approach provides superior efficiency in terms of parameter count and inference speed while maintaining competitive accuracy.
| Jurnal | ТАТУ хабарлари |
|---|---|
| Nashr sanasi | 2025-10-22 |
| DOI | 10.61663/253tuitmct3 |
DOI: 10.61663/253tuitmct3 · Maqolaning asl sahifasi · PDF
Image processing and computer vision place great importance on image segmentation, as dividing an image into different regions helps in better understanding its content.Image segmentation may involve several stages.This…
This paper investigates the use of Natural Language Processing (NLP) to monitor real-time public sentiment about Tashkent's public transportation system through Uzbek (Latin and Cyrillic) and Russian social media…
In this paper the methods for improving communication and efficiency in ROS2-based robotic systems are discusses.It highlights the growing complexity of industrial automation and the need for optimized algorithms to…
Modern electronic resources and web applications are exposed to numerous cyber threats, including DDoS attacks, SQL injections, XSS attacks, Slowloris, and multi-vector attacks.This article presents an intelligent…
the qualitative properties of the heat diffusion equation ( ) ( )in a medium with double nonlinearity and exponentially varying density are studied.Theorems on global and asymptotic solutions are proved, and the…
The recent adoption of smart cameras in surveillance, smart homes and smart city facilities has cast serious questions on the issue of data protection and privacy.The standard security measures do not usually identify…
Today, the automatic recognition of emotions in human speech is gaining increased importance due to the growing role of speech interfaces in humancomputer interaction applications.In this context, the present study…
This paper presents an algorithm in order to identify adverbs and adverbials during translation from Uzbek to English for machine translation.It is irrefutable that the majority of people struggle with distinguishing…
this article presents a solution to a linear programming problem and a method that is important for optimizing parameter definition.