Detecting image similarity using SIFT.

Sri, Kurra Hima; Manasa, Guttikonda Tulasi; Reddy, Guntaka Greeshmanth; Bano, Shahana; Trinadh, Vempati Biswas

Abstract

Manually identifying similarity between any images is a difficult task. This study proposes an image similarity detection model. The scale-invariant feature transform (SIFT) algorithm is used to detect similarity between input images, and also to calculate the similarity score that defines the extent to which the images are similar. SIFT detects the keypoints and computes its descriptors. A FLANN-based algorithm is used to find the best matches of the descriptors, taking the descriptor of first image and comparing it with the second image. The model achieved 60% accuracy in translational image similarity and 90% in rotational image similarity; feature-matching similarity differed depending upon the given inputs.

Citation

SRI, K.H., MANASA, G.T., REDDY, G.G., BANO, S. and TRINADH, V.B. 2022. Detecting image similarity using SIFT. In Jacob, I.J., Gonzalez-Longatt, F.M., Shanmugam, S.K. and Izonin, I. (eds.) Proceedings of the 2021 International conference on expert clouds and applications (ICOECA 2021), 18-19 February 2021, Bangalore, India. Lecture notes in networks and systems, 209. Singapore: Springer [online], pages 561-575. Available from: https://doi.org/10.1007/978-981-16-2126-0_45

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 2021 International conference on expert clouds and applications (ICOECA 2021)
• Start Date: Feb 18, 2021
• End Date: Feb 19, 2021
• Acceptance Date: Dec 31, 2020
• Online Publication Date: Jul 16, 2021
• Publication Date: Dec 31, 2022
• Deposit Date: Jul 4, 2024
• Publicly Available Date: Jul 4, 2024
• Publisher: Springer
• Peer Reviewed: Peer Reviewed
• Pages: 561-575
• Series Title: Lecture notes in networks and systems
• Series Number: 209
• Series ISSN: 2367-3370; 2367-3389
• ISBN: 9789811621253
• DOI: https://doi.org/10.1007/978-981-16-2126-0_45
• Keywords: Similarity detection; Image processing; Machine learning; SIFT; Keypoints; Descriptors; FLANN; Matching; Similarity
• Public URL: https://rgu-repository.worktribe.com/output/2063988