Computing feature vectors of students for face recognition using Apache Spark

With a massive increase of images and videos in the internet, processing these data for face detection and recognition is a bottleneck in current situation. At least there is a lack of frameworks and tools to process lots of data efficiently and not time consuming. System proposed in this paper offers face recognition framework based on Apache Spark to analyze lots images and videos from Instagram. As a people need to be searched and recognized are university students, their feature vectors used for recognition initially computed and can be enlarged at any moment. Spark's Resilient Distributed Databases gives ability to compute all the intermediate data directly in memory making computations faster meanwhile keeping resilience property.