Face occlusion detection and recovery using fuzzy comeans

مقدم أطروحة جامعية

Hadi, wildan Jamil

الجامعة

الجامعة التكنولوجية

الكلية

-

القسم الأكاديمي

قسم علوم الحاسوب

دولة الجامعة

العراق

الدرجة العلمية

ماجستير

تاريخ الدرجة العلمية

2008

الملخص الإنجليزي

Face occlusion that often occurs in practical face recognition systems may incur performance degradation, which is refers to the problem in which an entire region of the face is unknown.

Thus a technique that can recover the occluded part of the face is important for these systems. Several methods of detecting, and removing occlusion from gray face images have been reported by different researches.

Most of these algorithms use Principle Component Analysis and example-based learning methods to generate an image without occlusion, But these algorithms require large number of training face images and require dimensionality reduction methods.

So that, a new algorithm for occlusion detection and recovery is presented here for color face image without the need for a large number of training face images and dimensionality reduction methods. the proposed algorithim consists of three parts: Face Detection, Occlusion Detection, and face Recovery.

The proposed algorithm for face detection achieves 93 % in detection face images with clean background, whereas detection of face image with complex background is about 80 %, and detection of face images with occlusion and pure illumination is about 83 %.

Occlusion detection is implemented using a two step fuzzy c-means clustering algorithm, finally.

a proposed algorithm for face occlusion recovery achieves 73 % recovery quality.

التخصصات الرئيسية

العلوم الهندسية والتكنولوجية (متداخلة التخصصات)

الموضوعات

• الإلكترونيات الرقمية

• APA
• MLA
• AMA

لغة النص

الإنجليزية

نوع البيانات

رسائل جامعية

رقم السجل

BIM-304985