Abstract: The So-called image copy detection based on an image and pasting in another location of the same image is common way to manipulate the image content and difficult to detect illegal copies of copyrighted images,. In this paper the existing system is based on matching pairs by descriptors mainly by using visual words for given image matching which is hard to distinguish between images. This technique is called Bag-Of-Words (BOW) Quantization which is used cannot solve the problem well. To address this problem Scale-Invarient Feature Transform (SIFT) matches between the images where the technique used is mainly based on BOW Quantization where global context regions is done to filter false images. And thus comparatively gives rich performance of encoding between the two images.
Keywords: Bag-Of-Words (BOW), Scale - Invarient Feature Transform (SIFT), Geometrical Consistency, Difference-of-Gaussians (DoG)
|
DOI:
10.17148/IJARCCE.2018.71206
