Enhancing Face Recognition Between Two Sibling Sisters Based on Convolutional Neural Networks

In the age of the Internet of everything, facial recognition technology has become a critical element for security and identity verification. Especially in the identification of biological siblings with high facial similarity, this technology faces enormous challenges. This study delves into the application of Convolutional Neural Networks (CNN) in this field, demonstrating its outstanding performance in complex and variable facial recognition tasks due to its advanced feature learning and abstraction capabilities. This study introduces data augmentation techniques to not only increase the data volume but also enrich its diversity. Consequently, the model can learn and adapt to a broader range of facial features and expressions. Furthermore, the optimization of regularization techniques and hyperparameter selection minimized the risk of overfitting and ensured the model’s generalization capability on unseen data. The experimental results are encouraging. With comprehensive adjustments to various hyperparameters and regularization techniques, on the validation set, the accuracy of the model was 95%. It also demonstrated 60% accuracy on a complex and challenging test set, validating its effectiveness and reliability in real-world application scenarios. This research not only proves the effectiveness of CNNs in handling challenging facial recognition tasks but also provides valuable insights and foundations for future studies and applications. It is believed that, with the further development and optimization of technology, the accuracy and application scope of facial recognition will be greatly expanded and enhanced.