Image forensics using color illumination, block and key point based approach.

Autor: Thakur, Abhishek, Jindal, Neeru
Předmět:
Zdroj: Multimedia Tools & Applications; Oct2018, Vol. 77 Issue 19, p26033-26053, 21p
Abstrakt: Every individual is keen to exhibit socialism and connectedness posting their personal photos and videos on several social websites. Thus, it has become literally easy for the onlookers to see and modify their photos and videos. Here the concept enumerates in picture as image forensics, whereby it is possible to examine the authenticity and genuineness of photograph and video into consideration. In addition to this, nowadays photograph and videos are considered as a firm and valid proof in the court room for investigation, validation and judgement. Several experts are continuously working in an image forensic field to discover and develop better techniques for the detection of forgeries in image and videos. Detection of image forgeries is done in two ways. Firstly the forged image we are already familiar with is called active forgery detection technique and secondly, where we don’t know the forgery, then is referred to as the passive forgery detection technique. Passive technique is incorporated to detect forgery in this paper where hybridization is used. We have used DWT, color illumination Algorithm, SLIC Algorithm; SIFT Algorithm, Correlation Coefficient Map generation Algorithm, Block Matching Threshold Algorithm and Feature Extraction Algorithm for the detection and ramifying forgeries. The novelty of the proposed hybrid technique is the use of color illumination which detect image edges and trace them correctly to detect forged region. We have tested 48 images from database and find out image forgery detection at image level with Precision = 97. 25%; Recall = 100% and F1 = 98. 53%. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index