Utilization of Database Spatial Queries for Region of Interest Extraction in Stored Images

Image processing, specifically extracting regions of interest, being a costly operation would benefit from a method of storing its output region and input image data in a database for future querying so as to prevent re-execution of the processing operation. This study investigates the performance of using built-in spatial queries found in a couple of modern Database systems to store and retrieve specific regions of image data. After testing two widely used Database software: MongoDB and Postgresql, results have shown that MongoDB performs better in terms of average and tail latency as well as efficiency in disk space consumption.