(For USM Staff/Student Only)

EngLib USM > Ω School of Electrical & Electronic Engineering >

Face recognition using llbp and bitwise-or llbp / Lee Shi Neng

Face recognition using llbp and bitwise-or llbp_Lee Shi Neng_E3_2011_NI
Projek ini adalah tentang perkembangan sistem pengecaman muka manusia daripada satu set pengkalan data imej. Dari pengkalan data muka, imej asal mempunyai saiz yang berbeza kerana imej-imej ini dikumpul dari kamera yang berbeza kualitinya. Cabaran yang dihadapi adalah gambargambar akan berbeza dari segi saiz, kecerahan, resolusi, dan juga pose subjek. Oleh sebab itu, semua gambar asal harus dipotong dan diubah ukurannya menjadi gambar yang sama dari segi ukuran, iaitu 100x100 piksel sebelum digunakan sebagai input untuk pemprosesan imej. Pemprosesan imej dilakukan dengan algoritma LLBP dan Bitwise-OR. Pada dasarnya, LLBP dan Bitwise-OR LLBP mempunyai operator yang sama, hanya cara operator LLBP dan BitwiseOR mendapatkan kod garis binari adalah berbeza. Imej output digunakan untuk pengenalan wajah. Pengenalan wajah dilakukan dengan menggunakan teknik PCA. Satu sampel gambar-gambar digunakan sebagai gambar latihan. Tahap pengenalan ditentukan dengan mencari nisbah antara jumlah gambar yang dikenali dan jumlah gambar yang dilatih. Set imej yang berbeza contohnya, imej-imej yang diambil dari jarak yang berlainan atau camera yang berbeza, boleh digunakan supaya kami dapat mencari faktor-faktor yang berpengaruh terhadap tahap pengenalan. Keputusan eskperimen menunjukkan LLBP mempunyai nisbah yang lebih tinggi. Hal ini bermakna Bitwise-OR adalah lebih terpengaruh terhadap pose dan jarak subject. ___________________________________________________________________________________ This project is about the development of a human face recognition system from a given set of database. From the facial database, raw face images are in different in sizes since they are collected from cameras of varying qualities. Challenges we face here are images captured are different in size, resolution, brightness and also pose of subject. Therefore, all raw images are cropped and resized into images of the same size, 100x100 pixels before they are used as input for image processing. Next, cropped images are used for face images processing. Face image processing is done by using LLBP and BitwiseOR algorithm. Basically, LLBP and Bitwise-OR LLBP have the same operator, just that the operator for LLBP obtains the line binary code separately while the operator for Bit-OR obtains the line binary code simultaneously. Output images are used for face recognition. Face recognition is done by using PCA technique. A sample of images is used as train images. Other images will be tested and recognized with reference to this train images. Recognition rate is obtained by finding the ratio between number of recognized images and number of trained images. Different set images, for example, images taken from different distance or different cameras, can be used so that we can find out factors that have significant influence on recognition rate. Based on experimental results, the LLBP outperformed Bitwise-OR. This shows that Bitwise-OR is more sensitive towards pose and distance of subject.
Contributor(s):
Lee Shi Neng - Author
Primary Item Type:
Final Year Project
Language:
English
Subject Keywords:
human face recognition ; raw face images ; cameras
First presented to the public:
4/1/2011
Original Publication Date:
3/4/2020
Previously Published By:
Universiti Sains Malaysia
Place Of Publication:
School of Electrical & Electronic Engineering
Citation:
Extents:
Number of Pages - 77
License Grantor / Date Granted:
  / ( View License )
Date Deposited
2020-03-04 11:39:31.521
Submitter:
Nor Hayati Ismail

All Versions

Thumbnail Name Version Created Date
Face recognition using llbp and bitwise-or llbp / Lee Shi Neng1 2020-03-04 11:39:31.521