Accelerating Polynomial-Based Image Secret Sharing Using Hadoop

Abstract

Polynomial-based secret sharing techniques are most preferred in many applications. Image secret sharing techniques based on polynomial functions become computationally heavy if image size is large. Sequential operations are not that effective for construction and reconstruction algorithms on large size images. Parallel computing works efficiently when load and time reduction on a single machine is required. The concurrent execution of polynomial operations on image pixel values while applying steps of construction of shares and reconstruction of the secret makes the scheme more efficient. Hadoop-based concurrent approach is proposed for polynomial-based image secret sharing scheme. The experimental results show that Hadoop-based cluster speeds up the process up to 18% by one slave and up to 14% by adding two slave nodes. The addition in number of slave nodes accelerates the performance up to 50% for very large size images in the polynomial-based secret sharing algorithms.