THE 3 rd INTERNATIONAL SCIENTIFIC CONFERENCES OF STUDENTS AND YOUNG RESEARCHERS dedicated to the 99
th
anniversary of the National Leader of Azerbaijan Heydar Aliyev
167
3. Detect face boundaries and face landmarks, which are the
unique
features
of each face.
4. Make
measurements
of specific parameters of faces, such as
distance between eyes, etc. and compare with the previously
trained data, using the Deep Convolutional Neural Network
(DCNN) and support vector machine classifier (SVM)
algorithms.
Design Architecture The attendance checking system will perform this face recognition
algorithm to detect the attendees and will output the attendance list. The
system architecture has been shown in the Figure 2.
Figure 2. Attendance checking system architecture
With a camera, a person’s image is captured, and the face
recognition system detects face landmarks, encodes to the face
measurements, and compares the results
with
the database. If the
matching face is found in the database, user’s name is written in the
csv file together with the joining time.
Methodology To achieve this system, Python is preferred. The required libraries
for that project are:
OpenCV-python: This library combines more than 10 modules, 2 of which are used in this system: Image Processing and Video I/O. 1. Image Processing module is used for linear/non-linear filtering,
geometric image transformation, color space conversion, etc.
2. Video I/O module is utilized for taking video/video-codecs.
Dlib: This one is written for C++ and combines several ML algorithms.
For example, we will use deep learning, SMO based regression, SURF,
HOG, FHOG, etc.
Face recognition: It is used to detect the patterns of faces.
