To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Human Face Recognition from Part of a Facial Image based on Image Stitching |
|
Author |
Osama R. Shahin, Rami Ayedi, Oussama Ghorbel |
| Citation |
Vol. 25 No. 11 pp. 71-78 |
|
Abstract |
In the human community, because of the high rates of crime and the multiplicity of methods, some public institutions and markets have planted surveillance cameras. These cameras utilized to photograph crimes and their perpetrators and help the police track down the offenders. From these images, the investigators are able to identify them. However, a major problem facing investigators is when only a part of the perpetrator’s face appears in the image, which makes it difficult to capture the culprits. This work proposes an improved algorithm for human face recognition by using two approaches Eigenfaces and geometrical methods from part of a facial image based on image stitching. Image stitching is the process during which distinctive photographic images combined to make a complete scene or a high-resolution image. Several images integrated to form a wide-angle panoramic image. The quality of image stitching determines by calculating the similarity among the stitched image and original images, and by the presence of the seam lines through the stitched images. The Eigenfaces approach utilizes PCA calculation to reduce the feature vector dimensions. It provides an effective approach to discovering the lower-dimensional space. In addition to enabling the proposed algorithm to recognize the face, it also ensures a fast and effective way of classifying faces. The phase of feature extraction was followed by the classifier phase. Displacement classifiers using square Euclidean and City-Block distances were used. The test results demonstrate that the proposed algorithm gave a recognition rate of around 95%. |
|
Keywords |
Face recognition, Image stitching, Principal Component Analysis, Eigenfaces Distance Classifier, Geometrical approach. |
|
URL |
