Sparse representations for image decomposition with occlusions

@article{Donahue1996SparseRF,
  title={Sparse representations for image decomposition with occlusions},
  author={Michael J. Donahue and D. Geiger and Tyng-Luh Liu and Robert A. Hummel},
  journal={Proceedings CVPR IEEE Computer Society Conference on Computer Vision and Pattern Recognition},
  year={1996},
  pages={7-12}
}
  • M. Donahue, D. Geiger, R. Hummel
  • Published 18 June 1996
  • Computer Science
  • Proceedings CVPR IEEE Computer Society Conference on Computer Vision and Pattern Recognition
We study the problem of how to detect "interesting objects" appeared in a given image, I. Our approach is to treat it as a function approximation problem based on an over-redundant basis, and also account for occlusions, where the basis superposition principle is no longer valid. Since the basis (a library of image templates) is over-redundant, there are infinitely many ways to decompose I. We are motivated to select a sparse/compact representation of I, and to account for occlusions and noise… 

Figures from this paper

Image Recognition with Occlusions
TLDR
This work identifies all local minima of a given image so that a global optimization is possible by visiting all of them and investigates a greedy “L p Matching Pursuit” strategy.
Sparse Representations for Image Decompositions
TLDR
This work is motivated to select a sparse/compact representation for I, and to account for occlusions and noise in the image, and presents a concave cost function criterion on the linear decomposition coefficients that satisfies the requirements.
Matching pursuit‐based shape representation and recognition using scale‐space
TLDR
An analytical low‐level representation of images, obtained by a decomposition process, namely the matching pursuit algorithm, is proposed as a new way of describing objects through a general continuous description using an affine invariant dictionary of basis function (BFs).
Class Representation and Image Retrieval with Non-Metric Distances
TLDR
It is shown that the distance between two images is not a good measure of how well one image can represent another in non-metric spaces, and it is suggested that atypical points may be more important in describing classes.
Matching pursuit-based shape description and recognition in active contours
TLDR
A new framework linking image segmentation and shape recognition processes in the image processing domain is studied, based on a generic framework using sparse representation algorithms, and three different ways to introduce the a priori information in the active contour, reflecting three different segmentation goals are proposed.
Classification with Nonmetric Distances: Image Retrieval and Class Representation
TLDR
It is shown that in nonmetric spaces, boundary points are less significant for capturing the structure of a class than in Euclidean spaces, and it is suggested that atypical points may be more important in describing classes.
Image Analysis and Computer Vision: 1996
TLDR
A bibliography of nearly 2150 references related to computer vision and image analysis, arranged by subject matter, is presented, covering topics including computational techniques; feature detection and segmentation; image and scene analysis; and motion.
Condensing image databases when retrieval is based on non-metric distances
TLDR
It is shown that the distance between two images is not a good measure of how well one image can represent another in non-metric spaces, and it is suggested that atypical points may be more important in describing classes.
Unified framework for fast exact and approximate search in dissimilarity spaces
TLDR
A similarity retrieval framework which incorporates both of the aspects of similarity retrieval into a single unified model and shows that for any dissimilarity measure, the “amount” of triangle inequality can be changed to obtain an approximate or full metric which can be used for MAM-based retrieval.
On Fast Non-metric Similarity Search by Metric Access Methods
TLDR
This paper shows the triangular inequality can be enforced for any semimetric (reflexive, non-negative and symmetric measure), resulting in a metric that preserves the original similarity orderings (retrieval effectiveness), and proposes the TriGen algorithm for turning any black-boxSemimetric into (approximated) metric.
...
...

References

SHOWING 1-10 OF 25 REFERENCES
Image Recognition with Occlusions
TLDR
This work identifies all local minima of a given image so that a global optimization is possible by visiting all of them and investigates a greedy “L p Matching Pursuit” strategy.
Matching pursuit of images
  • F. Bergeaud, S. Mallat
  • Computer Science
    Proceedings., International Conference on Image Processing
  • 1995
TLDR
Results proved that the elements of the Gabor elements are precisely localized on the edges of the images, and give a local decomposition as linear combinations of "textons" in the textured regions.
On the recognition of occluded shapes and generic faces using multiple-template expansion matching
  • J. Ben-Arie, K. R. Rao
  • Engineering
    Proceedings of IEEE Conference on Computer Vision and Pattern Recognition
  • 1993
TLDR
The single-template method is proven to be equivalent to restoration of undergraded images using the Wiener filter and optimizes a practically defined matching criterion called discriminative signal-to-noise ratio (DSNR).
Matching pursuits with time-frequency dictionaries
The authors introduce an algorithm, called matching pursuit, that decomposes any signal into a linear expansion of waveforms that are selected from a redundant dictionary of functions. These
Atomic Decomposition by Basis Pursuit
TLDR
Basis Pursuit (BP) is a principle for decomposing a signal into an "optimal" superposition of dictionary elements, where optimal means having the smallest l1 norm of coefficients among all such decompositions.
Eigenfaces for Recognition
TLDR
A near-real-time computer system that can locate and track a subject's head, and then recognize the person by comparing characteristics of the face to those of known individuals, and that is easy to implement using a neural network architecture.
What Is the Goal of Sensory Coding?
  • D. Field
  • Computer Science
    Neural Computation
  • 1994
TLDR
It is proposed that compact coding schemes are insufficient to account for the receptive field properties of cells in the mammalian visual pathway and suggested that natural scenes, to a first approximation, can be considered as a sum of self-similar local functions (the inverse of a wavelet).
On a conjecture of Huber concerning the convergence of projection pursuit regression
with am a minimizing direction [or as is demonstrated easily a maximizing direction for E(gm)2]. Huber establishes weak L2(P) convergence of the procedure [rm -O 0 weakly in L2(P)]. Also in the
Basis pursuit
  • Shaobing Chen, D. Donoho
  • Computer Science
    Proceedings of 1994 28th Asilomar Conference on Signals, Systems and Computers
  • 1994
TLDR
Basis pursuit is a technique for decomposing a signal into an "optimal" superposition of dictionary elements, which has several advantages over matching pursuit and best ortho basis, including super-resolution and stability.
Entropy-based algorithms for best basis selection
TLDR
Adapted waveform analysis uses a library of orthonormal bases and an efficiency functional to match a basis to a given signal or family of signals, and relies heavily on the remarkable orthogonality properties of the new libraries.
...
...