智慧型校園大樓門禁監控系統
| dc.contributor | 臺北市立教育大學資訊科學系;國立臺灣師範大學機電工程學系 | zh_tw |
| dc.contributor.author | 蔡俊明 | zh_tw |
| dc.contributor.author | 葉榮木 | zh_tw |
| dc.date.accessioned | 2014-10-30T09:36:14Z | |
| dc.date.available | 2014-10-30T09:36:14Z | |
| dc.date.issued | 2007-12-20 | zh_TW |
| dc.description.abstract | 傳統的校園大樓門禁系統是以磁卡、鑰匙或密碼為主，其缺點是必須隨身攜帶，容易造成遺失或忘記，人員管理上，也會有漏洞。由於科技的進步，影像辨識技術可以應用於門禁監控系統，減少上述問題的發生。所以，本文提出以智慧型的影像辨識技術為基礎，來做校園大樓門禁監控系統。本系統主要包括人臉偵測和人臉辨識，其中人臉偵測包括膚色偵測和人臉確認，在膚色偵測中，因人的膚色容易受亮度的影響，所以，首先在不同亮度下，分別訓練出不同亮度的CbCr值，來做膚色偵測；接著，近距離的人臉，利用人臉中有眼睛和嘴唇以及上方有頭髮三種特徵，來確認出人臉區域，遠距離的人臉，利用變異度和頭髮，來驗證出人臉；最後，在人臉辨識中，利用主成份分析法和灰關聯分析法，做人臉辨識。經由實驗結果顯示，動態人臉偵測率和人臉辨識率分別達97.0％和94.3%。 | zh_tw |
| dc.description.abstract | 膚色偵測;人臉偵測;人臉辨識;主成份分析法;灰關聯分析法 | en_US |
| dc.identifier | ntnulib_tp_E0402_02_041 | zh_TW |
| dc.identifier.uri | http://rportal.lib.ntnu.edu.tw/handle/20.500.12235/36932 | |
| dc.language | chi | zh_TW |
| dc.relation | 亞洲大學主辦。2007全國計算機會議，台灣，台中。 | zh_tw |
| dc.relation | 2007 National Computer Symposium (NCS 2007), Taichung County, Taiwan. | en_US |
| dc.subject.other | The traditional entrance guard system of campus buildings relies mainly on magnetic stripe cards | zh_tw |
| dc.subject.other | keys or password. Its shortcoming must be hand-carried and it is apt to cause losing and forgetting. Also | zh_tw |
| dc.subject.other | the personnel management will have loophole. Because of the progress of science and technology | zh_tw |
| dc.subject.other | the face recognition can be applied to entrance guard's surveillance systems to reduce above described problems. Thus | zh_tw |
| dc.subject.other | this paper proposes an entrance guard surveillance system based on intelligent image recognition method. This system includes face detection and face recognition. The face detection includes skin detection and face identification. In the skin detection | zh_tw |
| dc.subject.other | the human skin is easily affected by luminance. So | zh_tw |
| dc.subject.other | we train different CbCr values for different class's luminance to segment the skin color. Next | zh_tw |
| dc.subject.other | eye | zh_tw |
| dc.subject.other | lip | zh_tw |
| dc.subject.other | hair and variance features are used to identify the face area. Finally | zh_tw |
| dc.subject.other | the Principle Component Analysis (PCA) and Grey Relational Analysis (GRA) are used to recognize the faces. Experiment results show that the proposed method has good performance. The dynamic face detection rate and dynamic face recognition rate are up to 97.0% and 94.3% | zh_tw |
| dc.subject.other | respectively. | zh_tw |
| dc.subject.other | Skin detection | en_US |
| dc.subject.other | Face detection | en_US |
| dc.subject.other | Face recognition | en_US |
| dc.subject.other | Principle component analysis | en_US |
| dc.subject.other | Grey relational analysis | en_US |
| dc.title | 智慧型校園大樓門禁監控系統 | zh_tw |
| dc.title | An Entrance Guard Surveillance System for Intelligent Campus Buildings | en_US |