The Viola-Jones algorithm in OpenCV is efficient for detecting faces. The study is the accuracy of face detection using Viola-Jones on FCNN. The data is divided into training, testing, and validation sets. The FCNN model achieves high accuracy but suffers from overfitting. Techniques such as regularization and dropout can improve performance. The training duration is relatively short. FCNN machine learning model. The first layer is a hidden layer with 128 neurons and uses the ReLU (Rectified Linear Unit) activation function. The second layer is the output layer with ten drilled neurons showing excellent performance, with a training accuracy of 99.49% and a validation accuracy of 97.68%. This shows that the model successfully learned patterns from training data and applied them effectively to validation data.