I am doing a project where the use Hadoop to recognize faces in jpg files. In recognition of faces using the library faint, for this to happen i need to get the whole picture. How do I get the whole picture at once using the MapReduce?
In my prior experience with image recognition with hadoop, the images were either individual files stored in hdfs or elsewhere, or we stored sets of the in the value field of sequence files.
In this way each map call would work on the entirety of a single image.