Question

Project title: Automated attendance system using facial recognition system

Describe in detail about this title of project

Research Depth :

- Depth of literature review

- Feasibility of the proposed project

- Implementation plan, methods and strategies

- Incorporating previous comments

Answer 1

The first question arises why this system comes into existence and what is the need of this project, actually the current method of taking the attendence in the institutions is passing on attendance or make a roll calls and mark the attendece of the student. this process is time consuming and sometimes it results in disturbing the class, if we talk about the biometric system for the professors or employees at the institutions or organization this system serve one at a time, this system is reliable but it is time consuming process so why not ship to the automatic attendence system which works on the face recognition technique, the automatic attendence system using facial recognition will automatically mark the attendence of the given particular person by extracting the feature of the image in a classroom or at entry gates the system will automatically mark the attendence of the students, professors or employees, if the image of the face of any given person matches with the any of the face in the given database the system has ability to find out that person extract the feature of that given image and mark the attendece of the particular person. This system is widely used in the various areas such as security control, police control, forensic medicine and management of the attendece system.

The Face recognition system concentrate on one face at a time and understand that even if the face is turned in a strange direction or in a bad lighting, it is still same person the system has that much of ability to find out that particular person by extracting the various features of the image. this system determines various unique features of a face that can distinguish it from the face of the any other people. these features could be size eyes, nose, length of the face, size of lips, skin of color, when all these features of image is compared from the face of the people in the database which is already known the system automatically mark the attendence of a given particular person as a human being our brain is capable to do all of these automatically and instantaneously but to design a system , we have to use some component so for capturing image we take a camera as an input and a python programming language to extract the features of the image and to mark the attendence.

The main building of the project that is:

1. Image which is going to acquiring image of the person
2. Face Detection which is going to detect the face of the person.
3. Feature Extention which is extract the features of the image.
4. Face Recognition which is extracted features are assign sort to characters of certain image.
5. Verificaton & Identification which is combine all the data.

Let us discuss briefly each block form above diagram:

Image Block: The following process will take place:

1. Caputure Image - the image capture will take place with the help of camera.
2. Enhancement of Image - to improve the charaterstics of the image, which perform generally contrast changing which makes bright pixels to more bright pixels, dark pixels to more dark pixels. Whenever the image is capture in the frame, the image contains pixels which is generally take care by Enhancment of Image.the tools available for image enhancement are:
   1. special filtering
   2. Frequency domain filtering

In general we say Enhancement of Image will take care of Quality of Image.

Face Detection Block: Once the get the image on the frame will detect the face of the person from the image, determining whether the human faces appear in the given image and where the faces are located.

The face detection and image acquisition, the main switch come is face recognition technique, how the face technique is performed, how the given face face match with the original face that store in the database. there are certain technique developed to recognise the faces, there are:

1. Hawk Technique - it is histogram of oriented gradients describes the object of appearance shape within the image by distribution of the intensity over the gradiants basically every images is in the form of pixel and every pixel form of a group called as cell basically all hawk technique is used to find out the vectors of all over region that contain the magnitude and direction of particular area, for each cell histogram gradient and direction are different which calculated, for different image we get different vector and different gradient, from that we distinguish the given face is the space or not after matching with the database then how this technique is performed.
2. Convolutional neural network: the term convolutional means scanning of the image, it is basically consists of two parts:
   1. layer
   2. Pulling

it will pass through the filter and other activation layer to give the final output as machine identifies image in the form of pixel and said to the pixel is scanned thought the filter, the filter can take the number of types like call filter, color filter, edge filter. the filter has random values that pass over the image and convolve to get the particular pixel. the convoluted convert if in the form to give and passes through the filter layer to get desired output.

This technique come from the Eigenface which is derived from the Eigenvectors that use in human face recognisation techniques. the eignevectors are derived from the covariance matrix of probability distribution over high dimesional vectors.

Feature Extraction Block: In the block determine what is specialization in the image. it is done with the help of following ways:

1. Segmentation which will draw the segments of the image. it is divided into three main parts:
   1. Point Segmentation - this features fix some points of the human faces.
   2. Line Segmentation - this feature will do join the lines of points of the human faces.
   3. Edge Segmentation - this feature will do the examine the outline of the human faces.
2. Morphological image processing is deal with structure of the faces.
3. Dialation - it is a operation that grows of thickness object in a binary frame.
4. Erosion - it is operation which controls elements which provide binary conditions.

so, there will be some variable which is used to perform the erosion function. if you going to perform Dialation there will be scalar which referencing the operations like think & fat faces. the thick and thin faces not an independent object depends on each other. The Dialation just boarding or thickening of the features where as erosion is expansion and contraction of the features so we have differentiate each and every faces so in this case we use Dialation & Erosion process.

Face Recognition Block: after getting indept detail of face from the image then going to assign a fixed set of image parameters to a constant parameter, parameters recognized by divided into two areas:

1. Patterns which describe using quantitative descriptions such as length area and texture of the image.
2. Symbolic information such as string and also described by the properties and the relationship between the certain symbols.

Verificaton & Identification Block: The identification is the process of analyzing that the certain phase of the certain individual whose image has been captured by us and the data will be stored once the person arrives on the camera and the attendence will be marked, the attendence will be stored in the form of excel sheet.

for example, when the person arrive, camera will capture the image, face detection, feature extraction, face recognition and finally image will be identified.