(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

Fpga-based facial detection and tracking design

Fpga-based facial detection and tracking design / Ng Chuang Hou
Pada tahun kebelakangan ini, pengesanan muka digunakan secara meluas dalam pengiktirafan muka yang mengandui sebahagian besar masa pengiktirafan dan membawa kepada kekangan masa algoritma pengiktirafan. Oleh itu, meningkatkan kelajuan dan mengurangkan kerumitan algoritma pengesanan wajah telah menjadi isu yang sangat penting, terutamanya untuk sistem pengesanan muka masa nyata. Projek ini membentangkan pelaksanaan Pengesanan Wajah dan Reka Bentuk Pelacakan yang berasaskan FPGA. Pengesanan muka adalah salah satu istilah biasa dalam pengawasan video. Sistem komputer boleh dilatih oleh algoritma yang berbeza yang dapat mengenal pasti wajah manusia dalam video masa nyata. Algoritma Viola-Jones dicadangkan dalam projek ini. Aliran video masa nyata dipaparkan pada monitor VGA melalui Cyclone IV FPGA board DE2-115. Video input berasal dari Terasic Multi-Touch LCD dengan Modul Kamera (MTLC). Perisian Quartus Prime Lite digunakan untuk mengendalikan FPGA. Walau bagaimanapun, kecekapan algoritma tidak mencapai tahap yang memuaskan. Selain itu, algoritma tidak berjaya digunakan dalam strim video masa nyata. Objektif projek tidak dicapaikan. _______________________________________________________________________________________________________ In the recent years, face detection is widely applied in face recognition which occupies most of the recognition time and leads to a time constraint of the recognition algorithms. Therefore, enhancing speed and reducing complexity of face detection algorithms have become a very important issue, especially for real-time face detection systems. This project presents the implementation FPGA-based Facial Detection and Tracking Design. Face detection is one of the common terms in video surveillance. A computer system can be trained by different algorithm which can identify the human face in the real time video. Viola-Jones algorithm is proposed in this project. Real time video stream is displayed on a VGA monitor through Cyclone IV FPGA board DE2-115. The input video comes from the Terasic Capacitive Multi-Touch LCD with Camera Module(MTLC). The Quartus Prime Lite software is used to operate The FPGA. However, the efficiency of the algorithm does not reach satisfactory level. Furthermore, the algorithm is not successfully applied in real time video stream. The objectives of the project are not achieved.
Contributor(s):
Ng Chuang Hou - Author
Primary Item Type:
Final Year Project
Identifiers:
Accession Number : 875008594
Language:
English
Subject Keywords:
face; detection; algorithms
First presented to the public:
6/1/2019
Original Publication Date:
3/3/2020
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 46
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2020-03-03 09:53:40.849
Date Last Updated
2020-12-02 16:02:35.1
Submitter:
Mohd Jasnizam Mohd Salleh

All Versions

Thumbnail Name Version Created Date
Fpga-based facial detection and tracking design1 2020-03-03 09:53:40.849