Patch Matching with Polynomial Exponential Families and Projective Divergences
نویسندگان
چکیده
Given a query patch image, patch matching consists in finding similar patches in a target image. In pattern recognition, patch matching is a fundamental task that is time consuming, specially when zoom factors and symmetries are handled. The matching results heavily depend on the underlying notion of distances, or similarities, between patches. We present a method that consists in modeling patches by flexible statistical parametric distributions called polynomial exponential families (PEFs). PEFs model universally arbitrary smooth distributions, and yield a compact patch representation of complexity independent of the patch sizes. Since PEFs have computationally intractable normalization terms, we estimate PEFs with score matching, and consider a projective distance: the symmetrized γ-divergence. We demonstrate experimentally the performance of our patch matching system.
منابع مشابه
Designing Deterministic Polynomial-Space Algorithms by Color-Coding Multivariate Polynomials∗
In recent years, several powerful techniques have been developed to design randomized polynomial-space parameterized algorithms. In this paper, we introduce an enhancement of color coding to design deterministic polynomial-space parameterized algorithms. Our approach aims at reducing the number of random choices by exploiting the special structure of a solution. Using our approach, we derive po...
متن کاملIntrinsic Geometry on the Class of Probability Densities and Exponential Families
We present a way of thinking of exponential families as geodesic surfaces in the class of positive functions considered as a (multiplicative) sub-group G of the group G of all invertible elements in the algebra A of all complex bounded functions defined on a measurable space. For that we have to study a natural geometry on that algebra. The class D of densities with respect to a given measure w...
متن کاملOn Rényi and Tsallis entropies and divergences for exponential families
Many common probability distributions in statistics like the Gaussian, multinomial, Beta or Gamma distributions can be studied under the unified framework of exponential families. In this paper, we prove that both Rényi and Tsallis divergences of distributions belonging to the same exponential family admit a generic closed form expression. Furthermore, we show that Rényi and Tsallis entropies c...
متن کاملOn Bending Response of Doubly Curved Laminated Composite Shells Using Hybrid Refined Models
This paper presents a static analysis of laminated composite doubly-curved shells using refined kinematic models with polynomial and non-polynomial functions recently introduced in the literature. To be specific, Maclaurin, trigonometric, exponential and zig-zag functions are employed. The employed refined models are based on the equivalent single layer theories. A simply supported shell is sub...
متن کاملRelationship between Coefficients of Characteristic Polynomial and Matching Polynomial of Regular Graphs and its Applications
ABSTRACT. Suppose G is a graph, A(G) its adjacency matrix and f(G, x)=x^n+a_(n-1)x^(n-1)+... is the characteristic polynomial of G. The matching polynomial of G is defined as M(G, x) = x^n-m(G,1)x^(n-2) + ... where m(G,k) is the number of k-matchings in G. In this paper, we determine the relationship between 2k-th coefficient of characteristic polynomial, a_(2k), and k-th coefficient of matchin...
متن کامل