Mask Region Grow segmentation algorithm for low-computing devices

Abstract

In this paper, we present an efficient and accurate mobile-based mask region grow (MRG) algorithm for natural scene segmentation. The algorithm is divided into three modules: first the captured RGB image is transformed to L*a*b* color space, then a suitable mask is generated and finally MRG is applied. The proposed MRG is best suitable for segmenting multiple foreground objects of single type from complex background, as compared to other existing segmentation algorithms. For validation MRG is tested with different types of challenging datasets including natural plant leaf, flowers and other images available from Internet sources. It is found that MRG is one very fast and accurate segmentation algorithm.