EngLib USM > Ω School of Electrical & Electronic Engineering >

Stereo vision and image recognition /Khairul Anuar Norhassim

Stereo vision and image recognition_Khairul Anuar Norhassim_E3_2011_NI

Abstract Pemprosesan imej 3D telah banyak digunakan secara meluas sekali dalam industri dan perkhidmatan seperti perubatan, pengesan, dan banyak lagi. Projek ini lebih difokuskan kepada pengecaman muka. Projek ini mencadangkan sistem pengecaman muka yang menggunakan penglihatan stereo pasif untuk menangkap(mengambil) matlumat wajah tiga dimensi(3D) dan penyamaan 3D dengan menggunakan Penyamaan oleh Punca Perbezaan (Sum of Squared Difference, SSD). Sistem ini menggunakan sepasang kamera stereo untuk menempatkan dan mengecam muka manusia.Teknik ini meningkatkan cara pengecaman objek 2D dengan mempertimbangkan permukaan wajah 3D, di mana kestabilan relatif di bawah pelbagai keadaan pencahayaan. Pertama, muka-muka akan dikesan, kemudian dibina semula daripada imej stereo. Selepas itu, muka 3D dibina dengan menggabungkan data imej 2D dan kedalaman yang sesuai. Wajah 3D kemudiannya dikomposisi menjadi komponen-komponen utamanya. Komponen-konponen utama itu akan digunakan untuk mengecam wajah 3D dengan membandingkan ciri-ciri muka sekarang kepada muka yang disimpan di dalam tabung data. Hasilnya, pengecaman muka adalah efisyen and tepat. ___________________________________________________________________________________ 3D image processing is widely used in industries and services such as in medical, entertainment, detector, and many more. This project focuses more on face recognition. This project proposes a face recognition system that uses passive stereo vision to the captured three-dimensional (3D) facial information and 3D matching using a simple Sum of Squared Difference (SSD) algorithm. This system uses a stereo camera to locate and recognizes a person’s face. This algorithm improves the state-of-the-art monocular 2D object recognition techniques by additionally considering facial 3D surface, which is relative stable under different lighting conditions. First, faces are detected and their surfaces are reconstructed from the stereo images. Afterwards, a 3D face is composed by joining 2D image data and appropriate depth data. The 3D face is then decomposed into its principal component. The principal components are used to recognize a 3D face by comparing the characteristics of the current face to those of known individuals in a database. The result is an efficient and accurate face recognition algorithm.

Contributor(s): Khairul Anuar Norhassim - Author Primary Item Type: Final Year Project Language: English Subject Keywords: face recognition ; stereo camera ; image 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 - 80 License Grantor / Date Granted:   / ( View License ) Date Deposited 2020-03-04 14:59:04.924 Date Last Updated 2020-06-09 10:53:07.313 Submitter: Nor Hayati Ismail

All Versions

• Help | 
• About | 
• Privacy Policy