Distributed indexes design to accelerate similarity based images retrieval in airport video monitoring systems

Indexing metric spaces have long been a heated subject in the area of data processing. This paper proposes the method to accelerate similarity images retrieval in airport video monitor system via distributed indexing schema D-MVP (Distributed MVP-Tree), which divides the metric space into several pieces. Experiments on large high-dimensional metric spaces data show that D-MVP indexing model has more significant performance improvement compared with the traditional indexing structure and is highly scalable with the explosive growth of data.