SURF Algorithm with Convolutional Neural Network as Face Recognition Technique

Recent developments in technology need the methods to become more efficient in various conditions. We can see this situation is much visible in multimedia and verification, where biometrics are used to recognize somebody from images, both as detections of suspicious behaviors and user verification. The paper presents proposed technique for face verification by the use of hybrid method based on SURF and neural network classifier. On the input image SURF is searching for potential key points, for which it creates special maps. These descriptors are forwarded to neural network for analysis and verification. Proposed solution works well and experimental results show good efficiency.