Using Apache Spark Streaming and Kafka to Perform Face

Recognition on Live Video Streams of Pedestrians

Through the use of Apache Spark Streaming, Kafka, and OpenCV on a distributed platform, we propose a method for recognising pedestrians in real time. This method intends to recognise motion and faces by matching new photos to a database of previously identified images. Apache Spark Streaming and Kafka have been utilised as real time analysis frameworks, which deliver event based decision making on Hadoop's distributed environment, because video processing and analysis from multiple resources is time consuming when using the Cloud or even any single highly configured machine. If real-time event analysis can be performed continuously, a choice may be made either immediately thereafter or simultaneously with the occurrence of the event in question. Hadoop is the foundation of all real-time analysis distributed solutions, so even processing massive amounts of films in parallel is no longer a bottleneck. Whenever continuous monitoring and decision making is involved with a large data set, this event based analysis can be put to use. This includes, but is not limited to, the monitoring of border areas of countries by cameras and drones, road traffic monitoring, the life science domain, airlines, logo recognition, and more