Large-Scale Face Image Retrieval Based on Hadoop and Deep Learning

In this paper, we analyzed large-scale face image retrieval method in cloud computing environment. Based on Hadoop platform, a large-scale face retrieval method by using deep learning was established to take advantage of parallel computing of MapReduce, and thus improving the efficiency of facial image retrieval. This method uses Histogram of Oriented Gradients and adopts a deep residual network. Meanwhile, Hadoop was adopted to process data parallelly to improve the retrieval efficiency. CASIA WebFace dataset was chosen for testing the proposed method, and the results of the experiment could demonstrate the effectiveness of the new method.