Multiscale framework for affine invariant pattern recognition and registration

Introduction

In many computer vision applications, there is a need to capture the characteristics of the objects depicted so that they are invariant under different geometric transformations. Oftentimes the problems arise, because there are rather limited number of such features availible. The multiscale framework offers a new approach to multiply the number of affine invariant features provided by traditional methods. The same ideas are also applicable to affine image registration. For more detailed description of the multiscale framework, see this introduction.

Software

Multi-Scale Autoconvolution (MSA)
Spatial multiscale affine invariants (SMA)
Comparison based histograms
Ridgelet invariants
Multiscale autoconvolution histograms
Generalized affine invariant moments
MSA based convexity measure C_a

Publications

E.Rahtu, M.Salo, and J.Heikkilä,
A new efficient method for producing global affine invariants,
Proc. 13th International Conference on Image Analysis and Processing (ICIAP 2005), Lecture Notes in Computer Science 3617, Springer, 407-414.

E.Rahtu, M.Salo, and J.Heikkilä,
Affine invariant pattern recognition using multi-scale autoconvolution,
IEEE Transactions on Pattern Analysis and Machine Intelligence (2005) 27(6):908-918.

J.Kannala, E.Rahtu, M.Salo, and J.Heikkilä,
A new method for affine registration of images and point sets,
Proc. 14th Scandinavian Conference on Image Analysis (SCIA 2005), Lecture Notes in Computer Science 3540, Springer, 224-234.

J.Kannala, E.Rahtu, and J.Heikkilä,
Affine registration with Multi-Scale Autoconvolution,
Proc. International Conference on Image Processing (ICIP 2005), Genoa, 3:1064-1067.

A.Teynor, E.Rahtu, L.Setia, and H.Burkhardt,
Properties of patch based approaches for the recognition of visual object classes,
Proc. 28th Jahrestagung der Deutschen Arbeitsgemeinschaft für Mustererkennung (DAGM 2006),
Lecture Notes in Computer Science 4174, 284-293.

E.Rahtu, M.Salo, J.Heikkilä, and J.Flusser,
Generalized affine moment invariants for object recognition,
Proc. 18th International Conference on Pattern Recognition (ICPR 2006), Hong Kong, 2:634-637.

E.Rahtu, M.Salo, and J.Heikkilä,
A new affine invariant image transform based on ridgelets,
Proc. 16th British Machine Vision Conference (BMVC 2006), Edinburgh, 3:1059-1068.

E.Rahtu, M.Salo, and J.Heikkilä,
Multiscale autoconvolution histograms for affine invariant pattern recognition,
Proc. 16th British Machine Vision Conference (BMVC 2006), Edinburgh, 3:1039-1048.

E.Rahtu, M.Salo, and J.Heikkilä,
A new convexity measure based on a probabilistic interpretation of images,
IEEE Transactions on Pattern Analysis and Machine Intelligence (2006) 28(9):1501-1512.

E.Rahtu, M.Salo, and J.Heikkilä,
Nonlinear functionals in the construction of multiscale affine invariants,
Proc. 15th Scandinavian Conference on Image Analysis (SCIA 2007), Lecture Notes in Computer Science 4522, Springer, 482-491.