DocumentCode :
2176450
Title :
Blind Identification Method for Authenticity of Infinite Light Source Images
Author :
Chen, Haipeng ; Shen, Xuanjing ; Lv, Yingda
Author_Institution :
Coll. of Comput. Sci. & Technol., Jilin Univ., Changchun, China
fYear :
2010
fDate :
18-22 Aug. 2010
Firstpage :
131
Lastpage :
135
Abstract :
In this paper, a blind identification for authenticity of infinite light source images was proposed. The inconsistency in light source direction resulted from image forgery is used to identify image authenticity. According to the error function of the actual light intensity and calculated light intensity and the constraint function of infinite light source to light beam, Hestenes-Powell multiplier method is used to calculate the light source direction of different objects and their background in infinite light source images. The image authenticity is determined based on the consistency between the light source direction of the object and its background. Experimental results showed that the authenticity of infinite light source images could be identified by the inconsistency in light source direction and the detection rate can reach 83.7%. The performance in terms of detection rate and time complexity of the proposed blind identification is superior to that of the existed method.
Keywords :
computational complexity; image recognition; matrix algebra; object detection; security of data; Hestenes-Powell multiplier method; blind identification method; constraint function; detection rate; image forgery; infinite light source image authenticity; light beam; light intensity; time complexity; Algorithm design and analysis; Computer science; Light sources; Lighting; MATLAB; Pixel; Reflectivity; Hestenes-Powell multiplier method; image authenticity; light source direction;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Frontier of Computer Science and Technology (FCST), 2010 Fifth International Conference on
Conference_Location :
Changchun, Jilin Province
Print_ISBN :
978-1-4244-7779-1
Type :
conf
DOI :
10.1109/FCST.2010.90
Filename :
5577301
Link To Document :
بازگشت