INCOPRATING CUDA IN HADOOP IMAGE PROCESSSING INTERFACE FOR DISTRIBUTED IMAGE PROCESSING

Computer Engineering

This paper presents parallel processing based on integrated approach of Hadoop and CUDA for large scale image processing. It makes the use of high reliability, scalability and fault tolerance capability of Hadoop system and high computing power of CUDA for processing huge amount of images in highly efficient manner.so the main aim is to improve performance of image processing task by using features of both, Hadoop and CUDA and to overcome the problem that are occur while processing large no of images in customary sequential manner. The proposed model serves as a good candidate solution for both type of applications i.e. data intensive application and compute intensive application. As Hadoop performs well for data intensive application through the use of HDFS and CUDA serves best in case of compute intensive application, integration of both the framework provides faster execution for image processing task. Image storage is provided through Hadoop Distributed File System and Map and Reduce primitive of Hadoop Mapreduce will be performed using CUDA on GPU.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.