Security in the ATM through manual processes is the main problem. So a face recognition technique has been implemented which is used to provide additional security to the ATM centres from theft with the combination of an embedded system.

In this process a webcam is placed outside the ATM centres and MATLAB technology is used for facial recognition and pattern analysis and identifies the face of the person then release the door lock. Once the person enters into the ATM centre the door automatically closes and if the person tries to break the ATM machine the vibrations sensors sends the signal and door locks immediately by using a PIC controller. Then an automated SMS is send to bank and also to nearby police station.Advantages of proposed system:
The perfect vibration detection.
Using web camera to detect the face authentication.
Automatic door mechanism.
Automated door lock in emergency situation.
Automated SMS notifications

Digital image processing and embedded systems are the dominating field in research. Face recognition in image processing is the emerging field because of its huge applications. A digital image that focuses the face of the person using a camera is given as input. In order to attain high accuracy and faster processing time digital image is taken as input. Due to lack of security and huge amount of accidents happening in ATM centres we have developed this project to ensure safety of the ATM centres. In this proposed framework only after recognition of the face pattern door for the ATM centre will be opened. If there occurrence of any unauthorized process then the control signal is sent to microcontroller to lock the door automatically. Thus, the implementation of this process is done using embedded system by using PIC Microcontroller and vibration sensor. After processing the face detection the control is given to the microcontroller to operate the door mechanism, if the person is trying to perform any theft action then the vibration sensor provides signal to the microcontroller to look the door automatically and message will besend to the authorized agency about the theft.