IMAGE SIMILARITY CHECKING USING SSIM
Airish Tom, Sr. Elsin C. SH · Zenodo (CERN European Organization for Nuclear Research) · 2023
Abstract— The Study focuses on developing and implementing the Structural Similarity Index Measure (SSIM) for automated image comparison tasks. The SSIM method is widely used to assess how similar the two images are to one another. Here, we use SSIM to find the similarity of paintings uploaded by the artists. The similarity is predicated on the notion that the structural information, brightness, and contrast of two similar images should be identical. By comparing the mean, standard deviation, and covariance of the pixels in the two images, SSIM calculates these variables. The code in this study compares an input image to a set of reference images by using the SSIM approach. Prior to converting both images to grayscale, the algorithm resizes the input image to match the size of the reference images. The highest score is then returned for each pair of images once the SSIM score is determined for each pair of images.