<![CDATA[During the pandemic, many activities that were originally carried out face-to-face were turned into virtual face-to-face due to reducing the spread of the virus. In conducting face-to-face activities, video conferencing applications are widely used for meetings and conducting activities such as learning. To support learning, sometimes teachers or lecturers find it difficult to observe whether the participants understand or not. In this study, used a dataset from KDEF which has seven classes. To find out the expressions of the participants, a model was made using the Convolution Neural Network that can detect human facial expressions. This convolution neural network has a contribution where the model used is a YOLO-face which is pipelined with CNN for classifications that have Alexnet architecture and modifications of Alexnet. The workings of this model is that the image is processed into a YOLO-face then the results from the YOLO-face are used by CNN classification to be able to classify the facial expressions of the participants. Then the photos will be classified using a CNN modification of the Alexnet architecture. The accuracy is 0.94 during training, precision is 0.92 during training, and recall is 0.96 during training. In this study, the face was successfully detected and classified the expressions on the face. For further development, it is necessary to optimize and increase the accuracy of the model to be able to classify facial expressions properly.]]>
Copyrights © 2021
