基于Hadoop的源相机识别方法研究

Source Camera Identification Using Hadoop

探索使用分布式方法解决数字图像取证中的关键问题——源相机识别问题的可能性。提出的解决方案是基于Lukas等人的算法,然后使用Hadoop框架中的MapReduce编程模型进行并行处理。使用几个优化来提高Lukas等人算法在Hadoop平台上的性能。初步试验结果显示,经过优化后的算法在Hadoop平台上的运行时间与单台计算机运行时间相比,分布式处理能够有效缩短任务的处理时间。

Explores the use of distributed methods to solve the key problem in digital image forensics-the possibility of source camera identification problems. The proposed solution is based on the algorithm of Lukas et al., and then uses the MapReduce programming model in the Hadoop framework for parallel processing. Several optimizations are used in this paper to improve the performance of the Lukas et al. algorithm on the Hadoop platform. The preliminary test results show that the distributed processing on the Hadoop platform can effectively shorten the processing time of the task compared with the single computer.