Optimized Image Regeneration for Old Damaged Reel Picture Using Novel Support Vector Machine Algorithm compared with K-Nearest Neighbors

This study will determine the sustainable development of old damaged reel photographs to improve regeneration through comparing the validation efficiency of the new SVM technique with K-Nearest Neighbors methods. The analysis considers a total of 20 accuracy samples of the Kaggle data set, divided into two sets of ten each to be analyzed. This therefore allowed us to deduce the accuracy values which were needed in establishing the significance of the image regeneration process. The G power value was set to a value of 0.8 and the SVM algorithm was observed to be at an accuracy of 89.54%. However, the accuracy of the KNN algorithm was 79.53%. The other results of our analysis included a 2-tailed significance value of 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> ) showing that the difference between the two algorithms was statistically significant. These results clearly state that, in terms of optimization of the accuracy of the process in the renewal of old ruined reel pictures, KNN performed worse than SVM. Besides showing the potential of the latest advances in ML based image restoration, this report also stresses the importance of adopting the appropriate algorithm to ensure that historical images are properly preserved. It is by that and others similar means that we can be sure that the priceless images of the visual records will not only be preserved but also restored for the sake of posterity.