A visualization platform for scaling-invariant boundary image matching

초록

In this paper, we present a visualization platform that supports the scaling invariance in boundary image matching. Once converting image boundaries into timeseries, we can use efficient time-series matching techniques and perform the boundary image matching fast even for large-scale image databases. In this boundary image matching, supporting the scaling invariance is very important to retrieve similar images much more correctly. In this paper, we design and implement the scaling-invariant boundary image matching as a client-server visualization platform. The client converts a query image into a time-series, sends it to the server with scaling ranges, and finally visualizes the returned results. The server performs the scaling-invariant matching efficiently on the boundary time-series database with multidimensional indexes. Experimental results show that our visualization platform performs the scaling-invariant matching very correctly by representing the matching results through various charts and graphs. © 2017 ICIC International.

키워드