Sei sulla pagina 1di 1

Solution

For identifying criminals from individual, we are developing an application which can be used
by Police and other law and enforcement officers. In this application the mobile phone
camera is used to scan the suspects face and if the suspect is criminal as per the police
records it will list whether he/she is a criminal and his/her criminal activities will be briefly
listed.
The application is based on deep Face Recognition (F R) emerged from deep learning.
Deep face recognition applies multiple processing layers to learn representation of data with
multiple levels of feature extraction. Police officers have to get the photographs of criminals
and this image is used for the learning process. We are using CNN (Convolution Neural
Network) for learning process. Convolutional neural networks (CNNs) are the most common
type of deep learning method for face recognition. The main advantage of deep learning
methods is that they can be trained with large amounts of data to learn a face representation
that is robust to the variations present in the training data. In this way, instead of designing
specialised features that are robust to different types of intra-class variations such as
illumination, pose, facial expression, age, etc. CNNs can learn them from training data.
First – a face detector is used to localize faces.
Second – The faces are aligned to normalise canonical coordinates.
Third – The face recognition (FR) module is implemented.
In FR module, face anti spoofing recognizes whether the face is live or spoofed. Face
processing is used to handle recognition difficulty before training and testing. The CNN
architecture VGG NET is used for our project. VGG Net has learned to extract the features
(feature extractor) that can distinguish the objects and is used to classify unseen objects.
VGG Net-16 has 99.52% accuracy and we are using it in our project. VGG Net architectures
and loss function are used to extract discriminative deep features when training. The
softmax loss is used as the supervision signal in objet recognition, and it encourages the
separability of features. Face matching method are used to do feature classification when
the deep features of testing data are extracted. The learning process are done in local
machine and deployed in cloud servers.
The main problem arise in the case of identical twins when one is not criminal and other one
is a criminal. Police can verify the authenticity by various biometric verification by taking
him/her into custody.

Potrebbero piacerti anche