Abstract | ||
---|---|---|
This paper presents a fast algorithm for retrieving plane geometry diagrams. Retrieving plane geometry diagrams has become an active research problem because that it is critical technology in various applications of searching the relevant materials of plane geometry such as search engine and auto tutorial system. Several existing systems use the features extracted from geometric diagram to retrieve the relevant materials of exercise problems of plane geometry, which achieve the better performance than the ones by using the keywords extracted from text. However, the existing retrieval methods do not consider the property of different features. This paper proposes a hierarchical strategy, which conducts the global-to-local bilayer retrieval. The first layer efficiently reduces the ranges by using global features; the second layer finds the relevant diagrams by using local features. This coarse-to-fine procedure can retrieve a short list of relevant diagrams with high similarity to the input diagram. This good performance lies in that the global features can describe the similarity of global structures and the local features can describe the delicate inner structure of diagrams. The experimental results demonstrate that the proposed algorithm can achieve good performance against the state-of-arts methods in terms of retrieval quality. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1145/3007669.3007671 | ICIMCS |
Keywords | Field | DocType |
Plane geometry diagram, diagram retrieval, hierarchical strategy, geometric primitives, shape detection | Search engine,Plane (geometry),Computer science,Geometric primitive,Theoretical computer science,Diagram | Conference |
Citations | PageRank | References |
0 | 0.34 | 1 |
Authors | ||
4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Wenbin Gan | 1 | 0 | 1.69 |
Xinguo Yu | 2 | 1 | 4.47 |
Sichao Lai | 3 | 0 | 0.34 |
Lei Xiang | 4 | 7 | 1.10 |