The Relevance of Non-generic Events in Scale Space Models
ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part I
Understanding and Modeling the Evolution of Critical Points under Gaussian Blurring
ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part I
Robust Real-Time Face Detection
International Journal of Computer Vision
On detecting all saddle points in 2D images
Pattern Recognition Letters
Exploring and exploiting the structure of saddle points in Gaussian scale space
Computer Vision and Image Understanding
Graph characteristics from the heat kernel trace
Pattern Recognition
IJCAI'83 Proceedings of the Eighth international joint conference on Artificial intelligence - Volume 2
Geometric characterization and clustering of graphs using heat kernel embeddings
Image and Vision Computing
Pattern analysis with graphs: Parallel work at Bern and York
Pattern Recognition Letters
A discrete scale space neighborhood for robust deep structure extraction
SSPR'12/SPR'12 Proceedings of the 2012 Joint IAPR international conference on Structural, Syntactic, and Statistical Pattern Recognition
Hi-index | 0.00 |
Linear or Gaussian scale space is a well known multi-scale representation for continuous signals. However, implementational issues arise, caused by discretization and quantization errors. In order to develop more robust scale space based algorithms, the discrete nature of computer processed signals has to be taken into account. Aiming at a computationally practicable implementation of the discrete scale space framework we used suitable neighborhoods, boundary conditions and sampling methods. In analogy to prevalent approaches, a discretized diffusion equation is derived from the continuous scale space axioms adapted to discrete two-dimensional images or signals, including requirements imposed by the chosen neighborhood and boundary condition. The resulting discrete scale space respects important topological invariants such as the Euler number, a key criterion for the successful implementation of algorithms operating on its deep structure. In this paper, relevant and promising properties of the discrete diffusion equation and the eigenvalue decomposition of its Laplacian kernel are discussed and a fast and robust sampling method is proposed. One of the properties leads to Laplacian eigenimages in scale space: Taking a reduced set of images can be considered as a way of applying a discrete Gaussian scale space.