結合臉部熱影像與雷射人體掃描特徵之身分辨識系統

An Identity Recognition System using Integration Face Thermal Image and Human Laser Scanned Features

林宏鎧(Hong-Kai Lin)

指導教授 ： 曾煥雯

國立臺灣師範大學/科技學院/工業教育學系/碩士(2013年)

摘要

由於科技進步迅速，許多利用傳統人力運作的產業，逐漸被自動化的機械、電腦系統所取代，在身分辨識領域內，以往需依靠人力進行門禁看守、檢查的工作，也逐漸被自動化的身分辨識系統取代。因此，如何使身分辨識系統能準確有效萃取並辨認他人人體特徵，即成很重要的關鍵。
本研究提出一套結合熱像儀萃取臉部溫度特徵與雷射掃描儀掃描影像萃取人體尺寸特徵，組成一套不易受光源、衣著影響的身分辨識系統。本系統分為資料庫建置與辨識共兩個階段，詳述如下。
於臉部溫度特徵資料庫建置，過程如下：首先將擷取的臉部熱影像，進行人臉鼻孔定位和半圓分割等前處理，利用溫度分層統計各溫度層的像素數量後，存入臉溫特徵資料庫；於雷射掃描人體尺寸特徵資料庫建置過程如下：先透過掃描儀取得雷射人體距離影像並進行前處理後，再藉由三角函數與肩膀定位，算出人體四項尺寸特徵，存入人體尺寸特徵資料庫。
系統在辨識階段，即利用人臉溫度和人體尺寸特徵資料庫的資料，配合實際量測的數值，進行人體尺寸特徵誤差，與臉溫卡方的參數計算後，將參數交由專家系統，透過內部的規則庫，進行身分的辨別。本研究樣本數為33人，經實驗分析得到的辨識率為96.9%。

關鍵字

身分辨識 ； 熱影像 ； 溫層分割 ； 雷射距離影像 ； 人體特徵

並列摘要

Nowadays, the development of biometric based identity recognition technology is getting faster, causing the industries to make transformation; especially the access management system, which are using fingerprint, voice, and iris recognition equipment to replace humans. As a result, how to make these identity recognition systems work efficiently and accurately, becomes the critical point.
　In this research, the researcher proposed an identity recognition system based on the fusion of thermal image and laser range image, and establish a human body feature database. The system is divided into two stages, which are the data build stage and recognition stage, as detailed below.
　In the part of thermal face feature database building stage, we use half-circle face segmentation and temperature layer splitter algorithm to do the feature extraction task. In the part of the human body feature database building stage, we use the human body range image and the trigonometric functions to calculate the height of the human body and the forehead as the basis of the identity recognition task.
　This system in recognition stage, we use the data of the face temperature and the body shape feature stored in the feature database, to calculate the error rate of the body shape and the chi-square value of the face temperature with the measured value, and use the expert system to do the recognition task. In the last part, we introduce a performance indicator and use 33 person experiment to get the 96.9% recognition rate.

並列關鍵字

identity recognition ； thermal image ； temperature layer splitter ； laser range image ； human body characteristics

參考文獻

[35] 李和謙（民100年），“應用溫度層技術之人臉辨識系統設計＂，國立臺灣師範大學工業教育學系碩士論文。

[4] Z. Lian, Z. Rui, Y. Chengbo, “Study on the Identity Authentication System on Finger Vein”, The 2nd International Conference on Bioinformatics and Biomedical Engineering, pp. 1905-1907, May 2008.

[20] S. Komandur, P. W. Johnson, R. L. Storch, and M. G. Yost, “Relation between Index Finger Width and Hand Width Anthropometric Measures,” 31st Annual International Conference of the IEEE EMBS Minneapolis, Minnesota, USA, September 2-6, 2009.

[1] P. Buddharaju, I. T. Pavlidis and P. Tsiamyrtzis, “Physiology-Based Face Recognition,” IEEE Conference on Advanced Video and Signal Based Surveillance, pp. 354-359, Sept. 2005.

[5] D. Zhang, W.K. Kong, J. You and M. Wong, “Onling Palmprint Identification,” IEEE Transactions on Instrumentation and Measurement, vol.25, pp. 1041-1050, 2003.