<title>MapSnapper: engineering an efficient algorithm for matching images of maps from mobile phones</title>

Abstract

The MapSnapper project aimed to develop a system for robust matching of low-quality images of a paper map taken from a mobile phone against a high quality digital raster representation of the same map. The paper presents a novel methodology for performing content-based image retrieval and object recognition from query images that have been degraded by noise and subjected to transformations through the imaging system. In addition the paper also provides an insight into the evaluation-driven development process that was used to incrementally improve the matching performance until the design specifications were met.

A tool for authoring unambiguous links from printed content to digital media

“…HotPaper [1] and Mobile Retriever [6] use features based on document text such as the spatial layout of words. Other systems such Bookmarkr [4] and MapSnapper [2] use pixel level image features, such as SIFT [7], to recognize generic document content. With these systems, visually obtrusive marks are not required for identification.…”

Embedded media markers

“…EMMs should not take up extra space on the paper, nor should the document layout be changed to accommodate the EMMs. 4. EMMs should minimize the semantic change to original paper content.…”

“…It is outdoor readable, foldable, and easy to share with other people [4,21]. However, it lacks dynamic location-specific information, such as video clip about a restaurant, weather forecast for a park, and currently available discounts for a hotel.…”

“…HotPaper [2] and Mobile Retriever [9] use features based on document text such as the spatial layout of words. Other systems such Bookmarkr [6] and MapSnapper [4] use pixel level image features, such as the SIFT [10] algorithm, to recognize generic document content such as pictures and graphic elements. With these systems, visually obtrusive marks are not required for identification.…”

Large-scale EMM identification based on geometry-constrained visual word correspondence voting

“…However, barcodes, when printed on a document for association with specific document contents, are visually obtrusive and could interfere with the document layout. Thus several current systems [11,12,13,17] rely on the document content itself for identification of media association. For example, from the camera-phone captured query image, SnapTell [12] and Kooba [13] extract visual features, and HOTPAPER [11] utilizes the layout of word boxes to determine the source page, and the location on the page for linkage to the corresponding media data.…”