FEATURE EXTRACTION IN HADOOP IMAGE PROCESSING INTERFACE

With the ease of Internet access and higher data rate availability to citizens all over the world, data sharing has become versatile in terms of type of data. Instead of just sharing data in textual form like few years back, people now are sharing the data in forms of pictures and videos. In most widely used social media platform – Facebook, users upload 350 million pictures every day. But, the traditional image processing systems are not capable of processing such a huge amount of data due to limitations in storage and computational capabilities. Moreover, MapReduce programming model of Hadoop framework provide processing of large amount of data. But since MapReduce programming model was created with intend to process the textual data, it is quite inefficient to process the images using MapReduce programming model directly. Also, technical complexity of MapReduce is high and hence before working on the image processing tasks, one has to understand this model. This is inconvenient and time consuming process. To ease this complexity, various image processing frameworks have been introduced which abstracts the MapReduce model while focusing on image processing task. Hadoop Image Processing Interface is one of those frameworks with various features and support to OpenCV.