Study of face recognition systems in light of environmental impact

碩士 === 國立臺北科技大學 === 建築系建築與都市設計碩士班(碩士在職專班) === 103 === Face image recognition rate, in addition to its own equipment conditions, is affected by ambient light environment very huge. The nature and the number of changes in light intensity, different variations of the illumination angle of light quantit...

Full description

Bibliographic Details
Main Authors: Huiju Chang, 張惠如
Other Authors: 周鼎金
Language:zh-TW
Published: 2015
Online Access:http://ndltd.ncl.edu.tw/handle/n88frh
Description
Summary:碩士 === 國立臺北科技大學 === 建築系建築與都市設計碩士班(碩士在職專班) === 103 === Face image recognition rate, in addition to its own equipment conditions, is affected by ambient light environment very huge. The nature and the number of changes in light intensity, different variations of the illumination angle of light quantity, can make between pixels in an image luminance distribution changes significantly, and may even generate a shadow in the face image. Because sometimes cast a shadow on the face image, the image directly will affect various statistical features extracted geometric features and characteristics of transformation, thereby affecting the face detection and recognition. This study is to verify the lobby of an international airport for the study. The main research method is based on field investigation and identifies the light source illuminating the impact of face recognition system problems and proposes countermeasures. By simulation that, to make the automatic face recognition clearance success rate of 90%, the most suitable environment lighting is the lighting environment of top light with smooth light.