In Sensus Futuris Facial Recognition System demonstration, first faces and their corresponding facial landmarks are detected, then a landmark based geometric normalisation is performed on each detected face. A deep network runs on the normalised face to extract salient information (face features). This information is useful for classifying faces of different people. It is robust to the variations in lighting and pose. The extracted features from the are matched against the faces in the watch list. When the top match is found with sufficient confidence, the name of the subject is presented. In live video demonstration, faces captured by the camera are presented on the screen with a bounding box around them. If the captured face is in our watch list, their name will be on the top of the box.