A Novel Approach for Digital Image Resemblance Using Deep Learning with Image Augmentation

Getting information on a picture is still a disorganized and time-consuming procedure in a society that seeks instant fixes for every issue. An individual must first determine the sort of image it is before doing a manual search for its data. In any online store, manually looking for a product and browsing through the entire website or application may take a lot of time, and there's a chance the user won't receive the outcome they were hoping for. Additionally, it might be challenging to grasp a product simply by reading it. In this study, we examine how well the models do in matching the photos to a large number of photographs in the dataset. Here, we look at the traits that the Convolutional Neural Network (CNN) model has, majorly VGGNet16. The dataset for the study was created entirely by us. The most significant part of picture resemblance was discovered to be feature extraction, data augmentation, and similarity measures. Furthermore, the results show that increasing the dataset increases accuracy even further.