AIR-IA: an analogous image removal approach using the intelligent archive
by Jyoti Malhotra; Jagdish Bakal
International Journal of Autonomic Computing (IJAC), Vol. 3, No. 3/4, 2020

Abstract: Deduplication is maturing as a standard attribute on backup and archives, whereby the aim is to free the storage space by removing the duplicates. Considering the storage room demand and justifiable deletion, this paper proposes a multi-container intelligent deduplication image archive system; where analogous images are disposed from the system based on the similarity approach. Similarity-aware image deduplication is achieved by calculating image fingerprints and the images are deleted when their hamming distance matches the predefined threshold. A probability model is addressed for the overall probability of getting similar images on the respective containers based on their relative storage and similarity scores of the images. In addition, the linear optimisation model is formulated to target data minimisation and storage space maximisation; which is further verified with the dataset. We perform experimentation of our work on the existing as well as synthesised datasets and various accuracy metrics are calculated in terms of precision, recall, f-score and deduplication ratio. It is observed that binary hashes used in our system give the fair contribution in removing similar images.

Online publication date: Tue, 20-Apr-2021

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Autonomic Computing (IJAC):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com