A markerless registration method for augmented reality based on affine properties

by: Y Pang, ML Yuan, AYC Nee, SK Ong, Kamal Youcef-Toumi

(2006), pp. 25-32.

View FullText article

Reviews [Write a review of this article]

There are no reviews of this article

Notes for this article

tryp has 0 private notes og 1 public note for this article.

Outlines a geometric registration technique based on affine reconstruction and reprojection properties using natural feature points identified by a KLT tracker. In two images, four non-coplanar feature points are selected as an affine origin and three affine basis points. A square of four points is then used to define a world coordinate system, which is then located in the 3d affine coordinate system. Affine reprojection matrices are then calculated for each frame, which are then used to reproject the world coordinate system. Camera pose is estimated using the plane used to define the world coordinate system. Methods of improving the robustness of this technique using feature point sets instead of individual points are then presented.

tryp (public ) - 2008-07-25 03:22:14

Find related articles from these CiteULike users

tryp

Find related articles with these CiteULike tags

3d, affine, model, tracking

@inproceedings{citeulike:3041541, address = {Darlinghurst, Australia, Australia}, author = {Pang, Y. and Yuan, M. L. and Nee, A. Y. C. and Ong, S. K. and Youcef-Toumi, Kamal }, booktitle = {AUIC '06: Proceedings of the 7th Australasian User interface conference}, citeulike-article-id = {3041541}, comment = {Outlines a geometric registration technique based on affine reconstruction and reprojection properties using natural feature points identified by a KLT tracker. In two images, four non-coplanar feature points are selected as an affine origin and three affine basis points. A square of four points is then used to define a world coordinate system, which is then located in the 3d affine coordinate system. Affine reprojection matrices are then calculated for each frame, which are then used to reproject the world coordinate system. Camera pose is estimated using the plane used to define the world coordinate system. Methods of improving the robustness of this technique using feature point sets instead of individual points are then presented.}, keywords = {3d, affine, model, tracking}, location = {Hobart, Australia}, pages = {25--32}, posted-at = {2008-07-25 03:22:14}, priority = {0}, publisher = {Australian Computer Society, Inc.}, title = {A markerless registration method for augmented reality based on affine properties}, url = {http://portal.acm.org/citation.cfm?id=1151758.1151760}, year = {2006} }

RIS record

TY - CONF ID - citeulike:3041541 L3 - citeulike-article-id:3041541 TI - A markerless registration method for augmented reality based on affine properties T2 - AUIC '06: Proceedings of the 7th Australasian User interface conference SP - 25 EP - 32 AD - Darlinghurst, Australia, Australia CY - Hobart, Australia PB - Australian Computer Society, Inc. KW - 3d KW - affine KW - model KW - tracking N1 - Outlines a geometric registration technique based on affine reconstruction and reprojection properties using natural feature points identified by a KLT tracker. In two images, four non-coplanar feature points are selected as an affine origin and three affine basis points. A square of four points is then used to define a world coordinate system, which is then located in the 3d affine coordinate system. Affine reprojection matrices are then calculated for each frame, which are then used to reproject the world coordinate system. Camera pose is estimated using the plane used to define the world coordinate system. Methods of improving the robustness of this technique using feature point sets instead of individual points are then presented. AU - Pang, Y AU - Yuan, ML AU - Nee, AYC AU - Ong, SK AU - Youcef-Toumi, Kamal PY - 2006/// UR - http://portal.acm.org/citation.cfm?id=1151758.1151760 ER -

RIS BibTeX