From the Publisher: 
The accessible presentation of this book gives both a general view of the entire computer vision enterprise and also offers sufficient detail to be able to build useful applications. Users learn techniques that have proven to be useful by first-hand experience and a wide range of mathematical methods. A CD-ROM with every copy of the text contains source code for programming practice, color images, and illustrative movies. Comprehensive and up-to-date, this book includes essential topics that either reflect practical significance or are of theoretical importance. Topics are discussed in substantial and increasing depth. Application surveys describe numerous important application areas such as image based rendering and digital libraries. Many important algorithms broken down and illustrated in pseudo code. Appropriate for use by engineers as a comprehensive reference to the computer vision enterprise.

Computer vision : a modern approach = 计算机视觉 : 一种现代的方法

Simulation and the monte carlo method

Local Shannon entropy measure with statistical tests for image randomness

Contrast in image processing is usually defined as a ratio between the darkest and the brightest spots of an image. In this paper we introduce a different contrast definition. The newly introduced Global Contrast Factor (GCF) corresponds closer to the human perception of contrast. GCF uses contrasts at various resolution levels in order to compute overall contrast. Experiments were conducted in order to find weight factors needed to calculate GCF. GCF measures richness of detail as perceived by a human observer, and as such can be used in various application areas like rendering, tone mapping, volume visualization, and lighting design.

/pdf/global-contrast-factor-a-new-approach-to-image-contrast-4soayayfc0.pdf

Global contrast factor - a new approach to image contrast

http://people.hws.edu/graham/Graham_VisionRes10.pdf

Statistical regularities in art: Relations with visual coding and perception

The aim of this paper is to study the visibility complexity ofdifferent regions in a 2D scene. Based on mutual information, which we used in our previous work to define scen e complexity, we propose two measures that quantify the complexity of a region from two different points of view.The knowledge of the complexity of a region can be useful to determine how difficult it is to recompute the visib ility links for an animation depending on the regions visited or to obtain the complexity of the movement of a robot . We also envisage its applicability to obtain an optimal load balancing in a parallel computation by dividin g the geometry in equal complexity regions.

/pdf/visibility-complexity-of-a-region-in-flatland-3zcg71chns.pdf

Visibility Complexity of a Region in Flatland

Ray tracing usually needs supersampling to reduce aliasingor noise in the final image. Not all the pixels in the image require the same quantity of rays, thusadaptive supersampling is implemented by adaptive subdivision of the sampling region, resulting in a refinement tree. We prese nt here a theoretically sound adaptive supersampling method based on entropy, an information theory approach wit h strong analogies to the decision tree problem where entropy is frequently used as a decision criterion. Ou r adaptive supersampling algorithm is implemented within a path tracing method and we show that our results comp are well to the ones obtained by a classic strategy.

/pdf/entropy-based-adaptive-supersampling-20o2l79u24.pdf

Entropy-based Adaptive Supersampling

We present a half-day course to review several information theory applications for computer graphics and visualization. Information theory tools, widely used in scientific fields such as engineering, physics, genetics and neuroscience, are also emerging as useful transversal tools in computer graphics and related fields. We introduce the basic concepts of information theory and how they map into application areas. Application areas in computer graphics include viewpoint selection, mesh saliency, scene exploration, ambient occlusion, geometry simplification, radiosity, adaptive ray-tracing and shape descriptors. Application areas in visualization are view selection for volume data, flow visualization, ambient occlusion, time-varying volume visualization, transfer function definition, time-varying volume visualization, iso-surface similarity maps and quality metrics. The applications fall broadly into two categories: the mapping of the problem to an information channel - as in viewpoint applications - and the direct use of measures such as entropy, Kullback-Leibler distance, Jensen-Shannon divergence, and f-divergences. These would be used to evaluate, for instance, the homogeneity of a set of samples being used as metrics. We will also discuss the potential applications of the information bottleneck method that allows us to progressively extract or merge information in a hierarchical structure.

Information theory in computer graphics and visualization

In this paper we investigate systematic sampling in the image- synthesis context. Systematic sampling has been widely used in stereology to improve the efficiency of different probes in experimental design. These designs are theoretically based on estimators of 1-dimensional and 2-dimensional integrals. For the particular case of the characteristic function, the variance of these estimators has been shown to be asymptotically N −−3/2, which improves on the O(N −−1) behaviour of independent estimators using uniform sampling. Thus, when no a priori knowledge of the integrand function is available, like in several image synthesis techniques, systematic sampling efficiently reduces the computational cost.

/pdf/systematic-sampling-in-image-synthesis-2u8vtik3ml.pdf

Systematic sampling in image-synthesis

In the radiosity method scene meshing has to accurately represent illumination variations avoiding unnecessary refinements that would increase the computational cost. In this paper we present three refinement criteria for hierarchical radiosity which are based on the information content of a ray between two patches or elements and the loss of information transfer between them due to the discretisation. The results obtained with our information-theory-based oracles improve on the classic ones based on transported power, kernel smoothness, and smoothness of received radiosity.

Jaume Rigau

Papers

Visibility Complexity of a Region in Flatland

Entropy-based Adaptive Supersampling

Information theory in computer graphics and visualization

Systematic sampling in image-synthesis

Information-theory-based oracles for hierarchical radiosity