## Radioptimization - Goal Based Rendering (1993)

Venue: | In Computer Graphics Proceedings, Annual Conference Series |

Citations: | 43 - 0 self |

### BibTeX

@INPROCEEDINGS{Kawai93radioptimization-,

author = {John K. Kawai and John K. Kawai and James S. Painter and James S. Painter and Michael F. Cohen and Michael F. Cohen},

title = {Radioptimization - Goal Based Rendering},

booktitle = {In Computer Graphics Proceedings, Annual Conference Series},

year = {1993},

pages = {147--154}

}

### Years of Citing Articles

### OpenURL

### Abstract

This paper presents a method for designing the illumination in an environment using optimization techniques applied to a radiosity based image synthesis system. An optimization of lighting parameters is performed based on user specified constraints and objectives for the illumination of the environment. The system solves for the "best" possible settings for: light source emissivities, element reflectivities, and spot light directionality parameters so that the design goals, suchastominimize energy or to give the the room an impression of privacy, are met. The system absorbs much of the burden for searching the design space allowing the user to focus on the goals of the illumination design rather than the intricate details of a complete lighting specification. A software implementation is described and some results of using the system are reported.

### Citations

1133 |
Practical Methods of Optimization
- Fletcher
- 1987
(Show Context)
Citation Context ...imization techniques, has resulted in a numberofvery useful texts under a number of headings, such as Mathematical Programming [6, 17], Operations Research [7], Optimal Control [19], and Optimization =-=[22, 8, 13]-=-. The important aspects of a particular optimization problem, leading to a choice of algorithm include: the nature of the objective function, e.g., linear vs. non-linear, convex vs. non-convex, di ere... |

747 |
Numerical Recipes
- Press, Flannery, et al.
- 1986
(Show Context)
Citation Context ...Newton's Method. These include the Davidon-Fletcher-Powell (DFP) and Broyden-Fletcher-Goldfarb-Shanno (BFGS) methods. These methods iteratively estimate the inverse Hessian from a series of gradients =-=[23, 22]-=-. In general, they begin with an identity matrix as the Hessian, thus defaulting to a simple gradient descent for the rst step and then modifying the inverse Hessian on succeeding iterations. A modi c... |

735 | The rendering equation
- Kajiya
- 1986
(Show Context)
Citation Context ...ysical constraints specify the relationships between light emission and element radiosities that are dictated by the physics of light transport. The constraints are captured in the rendering equation =-=[18]-=-. We assume perfect di use surfaces and a discretized environment yielding the radiosity approximation given in equation 2. Design goals are constraints provided by the user. These may be either equal... |

376 | A rapid hierarchical radiosity algorithm
- Hanrahan, Salzman, et al.
- 1991
(Show Context)
Citation Context ...best" possible settings for light source emissivities, surface re ectivities, and spot light directionality. Unconstrained optimization techniques are employed in conjunction with classical radio=-=sity [14, 4, 3, 16]-=- to simulate global illumination and our current implementation is thus limited to di use environments with xed geometry. Wehavehowever, extended the basic radiosity system to include spot lights as w... |

241 | A progressive refinement approach to fast radiosity image generation
- Cohen, Chen, et al.
- 1988
(Show Context)
Citation Context ...best” possible settings for light source emissivities, surface reflectivities, and spotlight directionality. Unconstrained optimization techniques are employed in conjunction with classical radiosity =-=[3, 2, 8]-=- to simulate global illumination. Creating an appropriate two-way link between the designer and the rendering system requires two important enhancements to basic rendering methods. First, since the de... |

132 |
The Gradient Projection Method for Nonlinear Programming, Part I. Linear Constraints
- Rosen
- 1960
(Show Context)
Citation Context ...aints are simple, and variables can be solved for as explicit expressions of other variables, then the rst alternative is attractive, as variables can be directly removed from the optimization problem=-=[24, 25]-=-. Other techniques introduce new variables as in the Lagrange Multiplier methods [22] in which the new unconstrained problem is taken to be the sum of the objective and a linear combination of the con... |

113 | Management of Large Amounts of Data in e o Interactive Building Walkthroughs
- Funkhouser, Sequin, et al.
- 1992
(Show Context)
Citation Context ...ingle room, architectural models. Clearly if the environment being modeled consists of many separate rooms, only a single local region should be considered. The work of Funkhouser, Sequin, and Teller =-=[12]-=- would be of great value here. In addition to the view independence, calculating perception in object space has the added advantage of faster performance if the number of elements is much smaller than... |

101 |
An efficient radiosity approach for realistic image synthesis
- Cohen, Greenberg, et al.
- 1986
(Show Context)
Citation Context ...ed from any view point without additional radiosity computations. A direct solution to the radiosity equation appears to require at least n 2 space and time, given n elements. Early radiosity methods =-=[5]-=- used substructuring techniques, decomposing the environment polygons into two levels of hierarchy to alleviate the problems of n 2 time and space. The shooting based progressive radiosity method [3] ... |

94 |
Applied Nonlinear Programming
- Himmelblau
- 1972
(Show Context)
Citation Context ...nstraints are allowed to be general non-linear functions. Research on optimization techniques, has resulted in a numberofvery useful texts under a number of headings, such as Mathematical Programming =-=[6, 17]-=-, Operations Research [7], Optimal Control [19], and Optimization [22, 8, 13]. The important aspects of a particular optimization problem, leading to a choice of algorithm include: the nature of the o... |

63 |
Globally Optimal Design
- Wilde
- 1978
(Show Context)
Citation Context ...imization techniques, has resulted in a numberofvery useful texts under a number of headings, such as Mathematical Programming [6, 17], Operations Research [7], Optimal Control [19], and Optimization =-=[22, 8, 13]-=-. The important aspects of a particular optimization problem, leading to a choice of algorithm include: the nature of the objective function, e.g., linear vs. non-linear, convex vs. non-convex, di ere... |

42 |
Lights from highlights and shadows
- Poulin, Fournier
- 1992
(Show Context)
Citation Context ...has been done on addressing the design problems in creating better quality lighting, except for a few systems that determine lighting placement by indicating desired areas of highlights and/or shadow =-=[12]-=-. Lighting designers base their art on the belief that spatial lighting patterns are a visual communicative medium, in which some patterns of light suggest or reinforce shared attitudes and impression... |

39 |
B.: Modeling the Interaction of Light Between Di use Surfaces. siggraph
- Goral, Torrance, et al.
- 1984
(Show Context)
Citation Context ...best" possible settings for light source emissivities, surface re ectivities, and spot light directionality. Unconstrained optimization techniques are employed in conjunction with classical radio=-=sity [14, 4, 3, 16]-=- to simulate global illumination and our current implementation is thus limited to di use environments with xed geometry. Wehavehowever, extended the basic radiosity system to include spot lights as w... |

30 |
Brightness function : effects of adaptation
- STEVENS, STEVENS
- 1963
(Show Context)
Citation Context ...s, limiting the amount of light entering the eye, in order to seek a state of equilibrium that is appropriate for the general brightness conditions. Tumblin and Rushmeier [18] studied work by Stevens =-=[17]-=- who theorized that the adaptation level of a scene can be estimated by the expected value (mean)sof the log10 of the luminances visible on the retina, i.e., EXPp∈retina{log10(L(p))} where L(p) is the... |

25 |
Introduction to Operations Research
- Ecker, Kupferschmid
- 1988
(Show Context)
Citation Context ...eneral non-linear functions. Research on optimization techniques, has resulted in a numberofvery useful texts under a number of headings, such as Mathematical Programming [6, 17], Operations Research =-=[7]-=-, Optimal Control [19], and Optimization [22, 8, 13]. The important aspects of a particular optimization problem, leading to a choice of algorithm include: the nature of the objective function, e.g., ... |

25 |
Numerical Methods for Constrained Optimization
- Gill, Murray
- 1974
(Show Context)
Citation Context ...imization techniques, has resulted in a numberofvery useful texts under a number of headings, such as Mathematical Programming [6, 17], Operations Research [7], Optimal Control [19], and Optimization =-=[22, 8, 13]-=-. The important aspects of a particular optimization problem, leading to a choice of algorithm include: the nature of the objective function, e.g., linear vs. non-linear, convex vs. non-convex, di ere... |

22 |
A Progressive Re nement Approach to Fast Radiosity Image Generation
- Cohen, Chen, et al.
- 1988
(Show Context)
Citation Context ...best" possible settings for light source emissivities, surface re ectivities, and spot light directionality. Unconstrained optimization techniques are employed in conjunction with classical radio=-=sity [14, 4, 3, 16]-=- to simulate global illumination and our current implementation is thus limited to di use environments with xed geometry. Wehavehowever, extended the basic radiosity system to include spot lights as w... |

20 |
Tone reproduction for realistic computer generated images
- Tumblin, Rushmeier
- 1991
(Show Context)
Citation Context ... asked to iteratively evaluate the visual impression from a rendered image, the images must provide (as much as possible) a subjective match to a “real” environment. The work of Tumblin and Rushmeier =-=[18]-=- on the psycho-physical quantities of subjective brightness has been applied to map luminance values to brightness values to provide higher fidelity for comparing the illumination of a computer genera... |

9 |
A study of subjective responses to low energy and nonuniform lighting systems. Lighting Design and Application
- Flynn
- 1977
(Show Context)
Citation Context ...of faster performance if the number of elements is much smaller than the number of screen pixels. 1.3.2 Subjective Impressions of Illumination In the 1970’s, John Flynn published a series of articles =-=[6, 4, 5]-=-, introducing a methodology with which to quantify parameters that elicit a shared human behavioral response and subjective impression. In particular, Flynn examined how non-uniform, peripheral, and b... |

9 |
Interim study of procedures for investigating the effect of light on impression and behavior" in Journal of the Illuminating Engineering Society, 3(2), 87-94. Psychology and Lighting Quality / 15 FGL, Fördergemeinschaft Gutes Licht
- Flynn, Spencer, et al.
- 1973
(Show Context)
Citation Context ...of faster performance if the number of elements is much smaller than the number of screen pixels. 1.3.2 Subjective Impressions of Illumination In the 1970’s, John Flynn published a series of articles =-=[6, 4, 5]-=-, introducing a methodology with which to quantify parameters that elicit a shared human behavioral response and subjective impression. In particular, Flynn examined how non-uniform, peripheral, and b... |

8 |
Tone reproduction for realistic computer generated images
- JackTumblin, Rushmeier
- 1991
(Show Context)
Citation Context ...eratively evaluate the visual impression from a rendered image, the images must provide (as much as possible) a subjective matchtoa\real"environment. We have applied the work of Tumblin and Rushm=-=eier [28]-=- on the psycho-physical quantities of subjective brightness as related to the adaptation of the viewer in order to map luminance values to brightness values for display on the CRT. This is an importan... |

7 |
A guide to methodology procedures for measuring subjective impressions in lighting, Journal of Illuminating Engineering Society, pp 95110
- Flynn, AIAFIES, et al.
- 1979
(Show Context)
Citation Context ...elief that spatial lighting patterns are a visual communicative medium, in which some patterns of light suggest or reinforce shared attitudes and impressions to people of the same cultural background =-=[10]-=-. In addition, the designer must be aware of the need to conserve the electrical energy used in implementing their designs. An over-reaction to the wasteful energy consumption of the 1960s and 1970s o... |

5 |
Nonlinear and Dynamic Programming
- Dano
- 1975
(Show Context)
Citation Context ...nstraints are allowed to be general non-linear functions. Research on optimization techniques, has resulted in a numberofvery useful texts under a number of headings, such as Mathematical Programming =-=[6, 17]-=-, Operations Research [7], Optimal Control [19], and Optimization [22, 8, 13]. The important aspects of a particular optimization problem, leading to a choice of algorithm include: the nature of the o... |

4 |
Brightness function: E ects of adaptation
- Stevens, Stevens
- 1963
(Show Context)
Citation Context ... to open and close, limiting the amount of light entering the eye toseekastate of equilibrium that is appropriate for general brightness conditions. Tumblin and Rushmeier [28] studied work by Stevens =-=[27]-=- who theorized that the adaptation level of a scene can be estimated by as the expected value (mean) of the log10 of the luminances visible on the retina: EXPp2retinaflog10(L(p))g (7) where L(p) is th... |

3 |
The hemi-cube: A radiosity for complex environments
- Cohen, Greenberg
- 1985
(Show Context)
Citation Context |

2 |
How to use INDSCAL: a computer program for canonical decomposition of N-way tables and individual di erences in multidimensional scaling
- Chang, Carroll
- 1972
(Show Context)
Citation Context ... data gathered resulted in a 6x6 symmetric dissimilarity matrix comparing the 6 rooms for each subject tested and each SD comparison made, e.g. large-small. A multidimensional scaling program INDSCAL =-=[2, 15]-=-, was used to determine how each subjected weighted the non-uniformity, peripheral and brightness values in making each SD comparison. A weighting of each dimension for each subject was determined tha... |

1 |
Teleological modeling. Siggraph Course Notes #30
- Barr
- 1989
(Show Context)
Citation Context ...m $150 to $275 per square foot [20], an extra investment of$1or $2 per square foot per year can potentially result in a large savings through improved productivity. This paper proposes a Teleological =-=[1]-=- or goal based illumination design approach to help a lighting designer search the space of of possible lighting speci cations. Though computers will never replace artists, the system may generate con... |

1 |
A study of subjective responsestolow energy and nonuniform lighting systems. Lighting Design and Application
- Flynn
- 1977
(Show Context)
Citation Context ...of faster performance if the number of elements is much smaller than the number of screen pixels. 2.3.2 Subjective Impressions of Illumination In the 1970's, John Flynn published a series of articles =-=[11, 9, 10]-=-, introducing a methodology with which to quantify parameters that elicit a shared human behavioral response and subjective impression. In particular, Flynn examined how non-uniform, peripheral, and b... |

1 |
Interim study of procedures for investigating the e ect of light on impression and behavior
- Flynn, Spencer, et al.
- 1973
(Show Context)
Citation Context ...of faster performance if the number of elements is much smaller than the number of screen pixels. 2.3.2 Subjective Impressions of Illumination In the 1970's, John Flynn published a series of articles =-=[11, 9, 10]-=-, introducing a methodology with which to quantify parameters that elicit a shared human behavioral response and subjective impression. In particular, Flynn examined how non-uniform, peripheral, and b... |

1 |
Optimal Control Theory. Prentice-Hall, Englewood Cli s
- Kirk
- 1970
(Show Context)
Citation Context ...ctions. Research on optimization techniques, has resulted in a numberofvery useful texts under a number of headings, such as Mathematical Programming [6, 17], Operations Research [7], Optimal Control =-=[19]-=-, and Optimization [22, 8, 13]. The important aspects of a particular optimization problem, leading to a choice of algorithm include: the nature of the objective function, e.g., linear vs. non-linear,... |

1 |
Model for visual respose to contrast
- Marimont
- 1962
(Show Context)
Citation Context ...of the residents. A better balance of goals between energy conservation and the quality of the lighting is needed [21]. With o ce and factory personnel costs ranging from $150 to $275 per square foot =-=[20]-=-, an extra investment of$1or $2 per square foot per year can potentially result in a large savings through improved productivity. This paper proposes a Teleological [1] or goal based illumination desi... |

1 |
Energy optimization and quality lighting design. Lighting Design and Application
- McKay
- 1986
(Show Context)
Citation Context ...s which were inadequately lit for their designed purposes, hampering the productivity of the residents. A better balance of goals between energy conservation and the quality of the lighting is needed =-=[21]-=-. With o ce and factory personnel costs ranging from $150 to $275 per square foot [20], an extra investment of$1or $2 per square foot per year can potentially result in a large savings through improve... |

1 |
Architectural Lighting for Commercial Interiors
- Sorcar
- 1987
(Show Context)
Citation Context ...designer a quick designcheck of what the room would look like. This also allows the customer who has no experience with lighting units a realistic preview of the nished room early in the design cycle =-=[26]-=-. Progress in rendering to date has mainly focused on improving the realism of the physical simulation and the development of algorithms with faster performance. Though great advances have been made i... |