Real-time eigenface-based face recognition system / Jeremias T. Lalis
By: Lalis, Jeremias T [author]
Description: 69 leaves : 29 cmContent type: text Media type: unmediated Carrier type: volumeSubject(s): Image processing -- Digital techniques | Image analysis -- Data processingDDC classification: 621.367 Dissertation note: Thesis (Master in Information Technology) -- Cebu Institute of Technology - University, March 2011. Summary: The system is composed of five main stages: First, and image of the face is acquired. This Acquisition is accomplished by using a webcam. Second, Viola-Jones method is employed to detect the location of the face in the acquired image and the hue histogram thresholding is used to enable the system to keep track of the face and its features. Third, once the system has targeted a face, the face image is normalized to make it less sensitive to lights variation, head scale and rotation. Fourth, the Principal Component Analysis (PCA) is used to analyze the spatial geometry of distinguishing features of the pre-processed/normalized face image. The final step is to let the eigenface to recognize the face image by looking for the training image that is closest to it in the PCA subspace using nearest-neighbor distance metrics. This is done by calculating Mahalanobis Distance, the distance from the projected test image to each projected training samples.| Item type | Current location | Home library | Call number | Status | Date due | Barcode | Item holds |
|---|---|---|---|---|---|---|---|
THESIS / DISSERTATION
|
GRADUATE LIBRARY | GRADUATE LIBRARY | 621.367 T L154 2011 (Browse shelf) | Not for loan | CL-T1755 |
Computer print-out.
Thesis (Master in Information Technology) -- Cebu Institute of Technology - University, March 2011.
The system is composed of five main stages: First, and image of the face is acquired. This Acquisition is accomplished by using a webcam. Second, Viola-Jones method is employed to detect the location of the face in the acquired image and the hue histogram thresholding is used to enable the system to keep track of the face and its features. Third, once the system has targeted a face, the face image is normalized to make it less sensitive to lights variation, head scale and rotation. Fourth, the Principal Component Analysis (PCA) is used to analyze the spatial geometry of distinguishing features of the pre-processed/normalized face image. The final step is to let the eigenface to recognize the face image by looking for the training image that is closest to it in the PCA subspace using nearest-neighbor distance metrics. This is done by calculating Mahalanobis Distance, the distance from the projected test image to each projected training samples.
There are no comments for this item.