Enhanced Image Forgery Detection Accuracy Using a SVM over Linear Regression

Image forgery detection is crucial in maintaining the integrity of digital content. This study compares the Novel Support Vector Machine algorithm with the Linear Regression algorithm to improve detection accuracy for image forgery. Both methods were tested with 20 samples each, and their parameters were adjusted to optimize performance. Using Gpower with an 80 % power level, a sample size of 20 was calculated for each group. The results indicate that the Novel Support Vector Machine achieved significantly higher accuracy (81.57 %) compared to Linear Regression (52.16 %). The statistical significance between the two methods was found to be 0.001 (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$p&lt;0.05$</tex>). Therefore, the SVM provides better results in predicting image forgery and enhances accuracy compared to the Linear Regression algorithm.