Swallow swarm fuzzy C-Means algorithm (SSFCM) and distributed processing framework for big image data

The recent period of image gathering process cannot be manipulated effectively on the system because of huge set of images and due to the high expenditure costs for the algorithms used to process on the recent images. Therefore, the processing of an image needs the help of distributed computing concepts. The subject named distributed computing is considered to be a complex area which requires sound technical learning and cannot be utilized by the research conducted so as to develop on the algorithms used for image processing; all the existing works are more concentrated on the optimization process used for processing the images with the aid of algorithms like-wise completely ignoring the inherent deficiency of an individual node-based processing method, where the quality of the image looks poor. In order to rectify this problem, static big image is processed by Image Cloud Processing (ICP) is a technique, which is proposed on modern days, which contains 2 different parts of processing methods i.e. SICP and DICP, so as to obtain the dynamic input of the static big image. The techniques used in the processing of small images are SICP and DICP that limits the capabilities of processing and the even images are broken down, at the time of cluster failure, when it needs to process the clusters on the timely basis. In order to rectify this problem, this paper proposes the design of Swallow Swarm Fuzzy C-Means Algorithm (SSFCM) based distributed processing framework for the clustering of images because of their poor requirements for computations. This work includes pre-processing of images with the help of image enhancement using Histogram Equalization (HE) and Adaptive Median Filtering (AMF). So as to accomplish SICP, 2 novel representations named Pure-image and Big-image are modeled to combine and execute by means of the algorithm named SSFCM-Map Reduce (SSFCM-MR) to attain large number of optimized configurations. In the formulated work, SSFCM algorithm is implemented through the Reduce Cluster to deal with the issues associated with clustering of a large image. DICP is implemented by means of a technique called as processing in parallel procedural working along with the traditional methods used for processing using the distributed system. The important feature of design in SSFCM-MR is the usage of affluent resources used for resources, given by the distributed system in the implementation of efficient parallel processing techniques. The output of the experiments of Corel-10k dataset are chosen to validate the potential of the formulated SSFCM-MR layout over the conventional state of arts techniques, in terms of efficiency and the standard of results.