Face Recognition Using Contour and Discrete Cosine Transform
Volume 14 Issue 2 2021
Download| Author(s): | Syed Muhammad Kashif Rafi , Mohsin Khan, Shahzad Nasim, Sheikh Muhammad Munaf |
|---|---|
| Abstract | Face recognition from human pose databases is a very famous work and many researchers add their interest in this field because of its diversity and working range. Biometric information processing have an impression on face recognition. It can be a vital component of natural human-machine interfaces. Face recognition using Contour and Discrete Cosine Transform (DCT) comprises to recognize the human face image from the database. In this research six hundred images of 200 persons are used with two different expressions. The images of database must be converted into 2-D format which is appropriate for the recognition of face because the recognition of color images are very typical. Two famous algorithm Geodesic distance transform for the extraction of geodesic distance paths and discrete cosine transform are used for the feature extraction process. The face was detected manually by taking the seed point on the nose of human face. For the extraction of paths fast marching algorithm is used. Affinity of Geodesic is intensity variation and this could extract more and more information from face images. This Geodesic distance also helped us to obtain the contours of the face. Contours are helpful to extracting the more exact features of images. In the end Discrete Cosine Transform (DCT) is applied to all the images that derived DCT coefficient. Euclidean distance for the classification of images are also used for the final analysis of images. The results are comparable to the state-of-the-art face recognition techniques. |
| Keywords | Weighted Distance Transform, Discrete Cosine Transform, Euclidean, Contour, 2D-images, KNN |
| Year | 2021 |
| Volume | 14 |
| Issue | 2 |
| Type | Research paper, manuscript, article |
| Journal Name | Journal of Information & Communication Technology (JICT) | Publisher Name | ILMA University | Jel Classification | - | DOI | - | ISSN no (E, Electronic) | 2075-7239 | ISSN no (P, Print) | 2415-0169 | Country | Pakistan | City | Karachi | Institution Type | University | Journal Type | Open Access | Manuscript Processing | Blind Peer Reviewed | Format | Paper Link | http://jict.ilmauniversity.edu.pk/arc/Vol14/14.2/3.pdf | Page | -- |