This paper is concerned with a vital topic in image processing:color image forgery detection. The development of computing capabilitieshas led to a breakthrough in hacking and forgery attacks on signal, image,and data...
详细信息
This paper is concerned with a vital topic in image processing:color image forgery detection. The development of computing capabilitieshas led to a breakthrough in hacking and forgery attacks on signal, image,and data communicated over networks. Hence, there is an urgent need fordeveloping efficient image forgery detection algorithms. Two main types offorgery are considered in this paper: splicing and copy-move. splicing isperformed by inserting a part of an image into another image. On the otherhand, copy-move forgery is performed by copying a part of the image intoanother position in the same image. The proposed approach for splicingdetection is based on the assumption that illumination between the originaland tampered images is different. To detect the difference between the originaland tampered images, the homomorphic transform separates the illuminationcomponent from the reflectance component. The illumination histogramderivative is used for detecting the difference in illumination, and henceforgery detection is accomplished. Prior to performing the forgery detectionprocess, some pre-processing techniques, including histogram equalization,histogram matching, high-pass filtering, homomorphic enhancement, andsingle image super-resolution, are introduced to reinforce the details andchanges between the original and embedded sections. The proposed approachfor copy-move forgery detection is performed with the Speeded Up RobustFeatures (SURF) algorithm, which extracts feature points and feature vectors. Searching for the copied partition is accomplished through matchingwith Euclidian distance and hierarchical clustering. In addition, some preprocessing methods are used with the SURF algorithm, such as histogramequalization and single-mage super-resolution. Simulation results proved thefeasibility and the robustness of the pre-processing step in homomorphicdetection and SURF detection algorithms for splicing and copy-move forgerydetection, respectively.
A geographically weighted regression (GWR) model with fewer explanatory variables and higher prediction accuracy is required in spatial analysis and other practical applications. This article proposes an l0-norm varia...
详细信息
A geographically weighted regression (GWR) model with fewer explanatory variables and higher prediction accuracy is required in spatial analysis and other practical applications. This article proposes an l0-norm variable adaptive selection method to enhance performances of a GWR by simultaneously performing model selection and coefficient optimization. Specifically, we formulate a regularized GWR model with an additional l0-norm constraint to shrink those unimportant regression coefficients toward zero and propose an adaptive variable selection algorithm by iteratively distinguishing the important variables from the variable set. At each location, the best variable subset and optimizing coefficient estimations are simultaneously achieved under the l(0)-GWR framework. Moreover, two novel criteria, the modified Bayesian information criterion and the interpretability of coefficient symbol, which specify the variable selection and model interpretation, respectively, are also introduced to improve the performance of the l(0)-GWR. Experiments on both simulated and actual data sets demonstrate that the proposed algorithm can significantly improve the estimation accuracy of coefficients and can also enhance the interpretative ability of the established model.
暂无评论