摘要：The classic use of Stream Processing platforms enablesworking with data in real time, which allows you togenerate data analysis quickly attending to a decisionmakingprocess. However, you can use these platformsfor other applications such as indexing and subsequentuse of similarity search objects in a database. The imagescan be displayed on a metric space, which has featuresthat allow rules to discard a not similar image quicklywithout making costly computations. This paper presentsthe use of a Stream Processing platform to index imagesgenerated by different users. For this, it is necessary torepresent these images by vectors containing differentMPGE-7 features. This paper shows a Stream Processingplatform using its processing elements (PEs) in parallel tospeed up the operations involved in the indexconstruction.
