Abstract

Exponentially growing digital images need good image analysis and comparison methods, which the study addresses. A selected collection of 10 photos from 2012 to 2023 is compared using powerful computer vision and feature extraction algorithms. Histogram of Oriented Gradients (HOG) is a strong and popular computer vision approach for extracting image information. This research uses HOG features to accurately depict visual structure. Extracting and measuring these traits leads to practical solutions for content-based image retrieval, image classification, and object identification. HOG characteristics from 10 images (2012–2023) are extracted and compared using cosine similarity in this research. Finding visually related photographs in the collection is simpler with the similarity matrix's entire view of the photos' visual correlations. HOG features and cosine similarity are unique in image similarity analysis, which is this paper's key contribution. Using image-based content retrieval and classification, this work improves computer vision and image processing