基於深度偵測之即時臉部辨識系統設計

Real-Time Face Recognition System Based on Depth Detection

周彥仰(Yen-Yang Chou)

指導教授 ： 陳永平

國立交通大學/電機學院/電機與控制工程系所/碩士(2015年)

https://doi.org/10.6842/NCTU.2015.00512

摘要

近年來，臉部辨識之應用逐漸受到重視，本論文提出結合人臉位置深度偵測之臉部辨識系統。整個系統以雙眼影像資訊為基礎，分為人臉位置深度偵測、臉部特徵擷取，以及臉部辨識三個部分。首先藉由臉部與其深度之偵測，決定目標是否進入系統設定之範圍，確認後即擷取臉部特徵，並進行臉部辨識以及身分認證。
本論文之主要貢獻在人臉深度偵測部分，提出一個利用雙眼資訊建構人臉深度之模型，用於快速深度偵測以判斷是否進行人臉辨識，並在臉部特徵擷取部分，結合了延伸局部二值模式、主成分分析與線性判別分析之方法，以改善臉部辨識之準確率。根據實驗結果，整個人臉深度偵測模型可以有效地偵測出目標之遠近，而臉部辨識在各人臉資料庫之準確率平均高達95%，且本系統可於0.2秒內完成人臉之認證，適用於即時人臉辨識系統。

關鍵字

深度偵測 ； 臉部辨識 ； 局部二值模式 ； 主成分分析 ； 線性判別分析

並列摘要

In recent years, the application of face recognition has received increasing attention. This thesis proposes a face recognition system based on depth detection. The system is based on binocular images and separated into three parts, including face and depth detection, facial feature extraction and face recognition. First, the face and depth detection are implemented to determine whether the object is getting into the region for face recognition. Once in the region, the system retrieves the facial features and processes the face recognition for identity verification.
This thesis mainly contributes to the construction of the depth model of an object’s face and to the extraction of facial features. The depth model is used to efficiently decide the object’s depth. And a combination of feature extraction method of extension of Local Binary Pattern, Principal Component Analysis and Linear Discriminant is applied to improve face recognition accuracy. From the experimental results, the depth model is efficient in detecting target’s depth; the recognition rate of proposed method achieves high recognition rate about 95% in several face databases, and the system is able to finish in 0.2 second which could be applied in real time system.

並列關鍵字

depth detection ； face recognition ； local binary pattern ； principal component analysis ； linear discriminant analysis

參考文獻

[1] Y. H. Kao, C. K. Liang, L. W. Chang, and H. H. Chen, “Depth Detection of Light Field,” in ICASSP 2007. IEEE International Conference on Acoustics, Speech and Signal Processing, 2007, pp. I-893-I-896.

[2] X. Fan, X. Wang, and Y. Xiao, “A shape-based stereo matching algorithm for binocular vision,” in 2014 IEEE International Conference on Security, Pattern Analysis, and Cybernetics (SPAC), 2014, pp. 70-74.

[3] A. Torralba and A. Oliva, “Depth estimation from image structure,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, pp. 1226-1238, 2002.

[6] T. Ojala, M. Pietikäinen, and D. Harwood, “A comparative study of texture measures with classification based on featured distributions,” Pattern Recognition, vol. 29, no. 1, pp. 51-59, 1996.

[7] T. Ahonen, A. Hadid, and M. Pietikäinen, “Face Description with Local Binary Patterns: Application to Face Recognition,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 28, pp. 2037-2041, 2006.