<![CDATA[This research aims to implement a Convolutional Neural Network (CNN) in facial expression classification using the Kaggle dataset which consists of five types of facial expressions, namely anger, disgust, fear, happiness and sadness. This method is considered important in supporting various applications such as emotion detection, facial recognition, and better human-machine communication. In this research, data preprocessing and augmentation were carried out using ImageDataGenerator to increase data diversity and prevent overfitting. Next, a CNN architecture is built which consists of convolution layers, pooling layers, and Dense layers. The model was trained using the Adam optimizer with a categorical crossentropy loss function for 50 epochs. The results show that the model achieves approximately 51% accuracy on the validation set. However, further analysis showed variations in model performance among facial expression classes, with some classes performing better than others. Keywords: Facial Expression Classification, Convolutional Neural Network, Kaggle Dataset, Data Augmentation, Image Processing.]]>
Copyrights © 2024
