Abstract

How best to efficiently establish correspondence among a large set of images or video frames is an interesting unanswered question. For large databases, the high computational cost of performing pair-wise image matching is a ma- jor problem. However, for many applications, images are inherently sparsely connected, and so current techniques try to correctly estimate small potentially matching subsets of databases upon which to perform expensive pair-wise match- ing. Our contribution is to pose the identi fication of potential matches as a link prediction problem in an image correspondence graph, and to propose an effec- tive algorithm to solve this problem. Our algorithm facilitates incremental image matching: initially, the match graph is very sparse, but it becomes dense as we al- ternate between link prediction and veri fication. We demonstrate the effectiveness of our algorithm by comparing it with several existing alternatives on large-scale databases. Our resulting match graph is useful for many different applications. As an example, we show the benefits of our graph construction method to a label propagation application which propagates user-provided sparse object labels to other instances of that object in large image collections.