## Design Galleries: A General Approach to Setting Parameters for Computer Graphics and Animation (1997)

### Cached

### Download Links

Citations: | 191 - 3 self |

### BibTeX

@MISC{Marks97designgalleries:,

author = {J. Marks and B. Andalman and P. A. Beardsley and W. Freeman and S. Gibson and J. Hodgins and T. Kang and B. Mirtich and W. Ruml and K. Ryall and J. Seims and S. Shieber},

title = {Design Galleries: A General Approach to Setting Parameters for Computer Graphics and Animation},

year = {1997}

}

### Years of Citing Articles

### OpenURL

### Abstract

Image rendering maps scene parameters to output pixel values; animation maps motion-control parameters to trajectory values. Because these mapping functions are usually multidimensional, nonlinear, and discontinuous, #nding input parameters that yield desirable output values is often a painful process of manual tweaking. Interactiveevolution and inverse design are two general methodologies for computer-assisted parameter setting in which the computer plays a prominent role. In this paper we present another such methodology.

### Citations

478 | Multilevel k-way partitioning scheme for irregular graphs
- Karypis, Kumar
- 1998
(Show Context)
Citation Context ...ptimal graph partitioning is NP-hard [4], but many good heuristics have been developed for this problem [1]. Our partitioning code is based on an algorithm and software developed by Karypis and Kumar =-=[9]-=-. Once the initial w-way partition is formed, representative lights for each partition are selected, and installed in the hierarchy. The partitioned subsets, minus their representative vertices, 1 Sin... |

388 |
Modern Multidimensional Scaling: Theory and Applications
- Borg, Groenen
- 2005
(Show Context)
Citation Context ...moved to the surrounding image gallery. Mousing on an image in the gallery highlights its associated thumbnail, and vice versa. Thumbnail layout is accomplished using a multidimensional scaling (MDS) =-=[3]-=- method due to Torgerson [24]. 4 Given a matrix of distances between points, MDS procedures compute an embedding of the points in a low-dimensional Euclidean space (2D in our case) such that the inter... |

378 | Some simplified NP-complete graph problems - Garey, Johnson, et al. - 1976 |

318 | Particle systems a technique for modeling a class of fuzzy objects
- REEVES
- 1983
(Show Context)
Citation Context ...3.3 are used. The hopper-dog DG is illustrated in the video proceedings. 4.3 Particle Systems Particle systems are useful for modeling a variety of phenomena such as re, clouds, water, and explosions =-=[16]-=-. A useful particle-system editor might have 40 or more parameters that the animator can set, so achieving desired e ects can be tedious. As in the previous subsection, we useaDGinterface to re ne an ... |

315 | Spacetime Constraints
- WITKIN, KASS
- 1988
(Show Context)
Citation Context ...ter settings, and the user acts as an objective-function oracle, interactively selecting computer-suggested alternatives for further exploration. A more automatic methodology is inverse design, e.g., =-=[10, 12, 14, 19, 22, 25, 27]-=-: the computer searches for parameter settings that optimize a user-supplied, mathematically stated objective function. Unfortunately, there are many interesting and important graphics processes for w... |

296 | Evolving Virtual Creatures
- Sims
- 1994
(Show Context)
Citation Context ...ter settings, and the user acts as an objective-function oracle, interactively selecting computer-suggested alternatives for further exploration. A more automatic methodology is inverse design, e.g., =-=[10, 12, 14, 19, 22, 25, 27]-=-: the computer searches for parameter settings that optimize a user-supplied, mathematically stated objective function. Unfortunately, there are many interesting and important graphics processes for w... |

251 |
WS: Theory & methods of scaling
- Torgerson
- 1958
(Show Context)
Citation Context ...ge gallery. Mousing on an image in the gallery highlights its associated thumbnail, and vice versa. Thumbnail layout is accomplished using a multidimensional scaling (MDS) [3] method due to Torgerson =-=[24]-=-. 4 Given a matrix of distances between points, MDS procedures compute an embedding of the points in a low-dimensional Euclidean space (2D in our case) such that the interpoint distances in the embedd... |

238 | Artificial Evolution for Computer Graphics - Sims - 1991 |

198 | Recent directions in netlist partitioning: A survey
- Alpert, Kahng
- 1995
(Show Context)
Citation Context ... cut set, the total cost of all edges that connect vertices in di erent subsets, is minimized. Optimal graph partitioning is NP-hard [4], but many good heuristics have been developed for this problem =-=[1]-=-. Our partitioning code is based on an algorithm and software developed by Karypis and Kumar [9]. Once the initial w-way partition is formed, representative lights for each partition are selected, and... |

147 |
Some distance properties of latent root and vector methods used in multivariate analysis Biometrika 53 325–338
- Gower
- 1966
(Show Context)
Citation Context ...true for the applications we discuss here, although it need not be true in general), classical scaling is equivalent to an e cient technique for computing a principal-component analysis of the points =-=[5, 13]-=-. The layouts computed by classical scaling are not without anomalies | as we are using it, this MDS method is a projection from a high-dimensional space onto a 2D space, which cannot be done without ... |

141 | Sequential updating of projective and affine structure from motion - Beardsley, Zisserman, et al. - 1994 |

123 |
Evolutionary Art and Computers
- Todd, Latham
- 1994
(Show Context)
Citation Context ...modeling, and motion-control processes. The notion of having the computer assist actively in setting parameters is therefore appealing. One such computer-assisted methodology is interactive evolution =-=[11, 21, 23]-=-: the computer explores the space of possible parameter settings, and the user acts as an objective-function oracle, interactively selecting computer-suggested alternatives for further exploration. A ... |

113 | Hierarchical Spacetime Control
- LIU, GORTLER, et al.
- 1994
(Show Context)
Citation Context ...ter settings, and the user acts as an objective-function oracle, interactively selecting computer-suggested alternatives for further exploration. A more automatic methodology is inverse design, e.g., =-=[10, 12, 14, 19, 22, 25, 27]-=-: the computer searches for parameter settings that optimize a user-supplied, mathematically stated objective function. Unfortunately, there are many interesting and important graphics processes for w... |

95 | Spacetime constraints revisited
- Ngo, Marks
- 1993
(Show Context)
Citation Context |

95 | Sensor-Actuator Networks
- Panne, Fiume
- 1993
(Show Context)
Citation Context |

92 | Generation of transfer functions with stochastic search techniques
- He, Hong, et al.
- 1996
(Show Context)
Citation Context ...s for two data sets: the simulated electron density of a protein, and a CT scan of a human pelvis. 3 The application of both interactive evolution and inverse design to this problem is the subject of =-=[7]-=-. MERL-TR-97-14 October 1997 4s3.1 Input and Output Vectors The protein data set contains values in the interval [0; 255]. The opacity transfer function over this domain is parameterized by a polyline... |

80 |
Efficient Calculation of Primary Images from a Set of Images," tEEE Transactions on Pattern Analysis and
- Murakami, Kumar
- 1982
(Show Context)
Citation Context ...true for the applications we discuss here, although it need not be true in general), classical scaling is equivalent to an e cient technique for computing a principal-component analysis of the points =-=[5, 13]-=-. The layouts computed by classical scaling are not without anomalies | as we are using it, this MDS method is a projection from a high-dimensional space onto a 2D space, which cannot be done without ... |

65 | The earth mover’s distance, multi-dimensional scaling, and color-based image retrieval
- Rubner, Guibas, et al.
- 1997
(Show Context)
Citation Context ...er t = 500; 000. To reach this point requires 8 500; 000 = 4; 000; 000 raycast 4 The use of more sophisticated MDS techniques for arranging a database of images is being investigated by Rubner et al. =-=[18]-=-. MERL-TR-97-14 October 1997 6soperations and takes less than 40 minutes on a single MIPS R10000 processor. This duration is roughly one-sixth of that needed to render the 256 full-size images (300 30... |

57 |
Painting with light
- Schoeneman, Dorsey, et al.
- 1993
(Show Context)
Citation Context |

42 |
Lights from highlights and shadows
- POULIN, FOURNIER
- 1992
(Show Context)
Citation Context ...cursor to image rendering. Previous attempts at computer-assisted lighting speci cation have used inverse design. For example, the user can specify the location of highlights and shadows in the image =-=[15]-=-, pixel intensities [19], or subjective impressions of illumination [10]; the computer then attempts to determine lighting parameters that best meet the given objectives, using geometric [15] or optim... |

21 |
High performance multibody simulations via symbolic equation manipulation and Kane’s method
- Rosenthal, Sherman
- 1986
(Show Context)
Citation Context ...ired forward velocity and hopping height, as well as desired positions for joints in some of the appendages. The equations of motion for the system are generated using a commercially available package=-=[17]-=-; dynamic simulation is used to produce the animations. We started with a basic hopping motion, and then used a DG approach to explore seven input quantities in order to achievestylistic, physically a... |

13 |
Some simpli ed NP-complete graph problems
- Gary, Johnson, et al.
- 1976
(Show Context)
Citation Context ...joint vertex subsets of equal cardinality such that the cost of the cut set, the total cost of all edges that connect vertices in di erent subsets, is minimized. Optimal graph partitioning is NP-hard =-=[4]-=-, but many good heuristics have been developed for this problem [1]. Our partitioning code is based on an algorithm and software developed by Karypis and Kumar [9]. Once the initial w-way partition is... |

11 |
Disney meets Darwin { the evolution of funny animated gures
- Ventrella
- 1995
(Show Context)
Citation Context ...age angular velocity, the number of velocity reversals, and the number of revolutions for each rod. Euclidean distance is used as the distance metric on this output space. 5Both interactive evolution =-=[26]-=- and inverse design [12, 14, 22, 25, 27] havebeen applied previously to motion control. 6Even without the application of external torques at its joints, the 2D double pendulum exhibits chaotic behavio... |

8 |
Pixel cinematography: a lighting approach for computer graphics
- Kahrs, Calahan, et al.
- 1996
(Show Context)
Citation Context ...e desired illuminative characteristics of the resulting image. This requirementmay be satis able in an architectural context [10], but seems very challenging in a more general cinematographic context =-=[8]-=-. The most di cult lighting parameters to set are those relating to light type and placement, so they have been the focus of our e orts. 2.1 Input and Output Vectors For the light selection and placem... |

8 |
Chaos in a double pendulum
- Shinbrot, Grebogi, et al.
- 1992
(Show Context)
Citation Context ...nd inverse design [12, 14, 22, 25, 27] havebeen applied previously to motion control. 6Even without the application of external torques at its joints, the 2D double pendulum exhibits chaotic behavior =-=[20]-=-. MERL-TR-97-14 October 1997 7sThe mapping from input vector to output vector is accomplished by dynamically simulating 20 seconds of the pendulum's motion, and using the algorithm in Figure 4 for dis... |

7 |
Sequential updating of projective and a ne structure from motion
- Beardsley, Zisserman, et al.
- 1997
(Show Context)
Citation Context ... lower right). A point- and line-based 3D model is extracted from a triplet of scene images, each taken from a di erent viewpoint. This reconstruction process is completely automatic, as described in =-=[2]-=-. Points and lines are then aggregated semi-automatically into planes. An illumination of the nal recovered model is used to modulate intensity in one of the original photographs. 3 Opacity and Color ... |

5 |
Radioptimization { goal-based rendering
- Kawai, Painter, et al.
- 1993
(Show Context)
Citation Context |

4 |
A prototype system for design automation via the browsing paradigm
- Kochhar
- 1990
(Show Context)
Citation Context ...modeling, and motion-control processes. The notion of having the computer assist actively in setting parameters is therefore appealing. One such computer-assisted methodology is interactive evolution =-=[11, 21, 23]-=-: the computer explores the space of possible parameter settings, and the user acts as an objective-function oracle, interactively selecting computer-suggested alternatives for further exploration. A ... |

2 |
Arti cial evolution for computer graphics
- Sims
- 1991
(Show Context)
Citation Context ...modeling, and motion-control processes. The notion of having the computer assist actively in setting parameters is therefore appealing. One such computer-assisted methodology is interactive evolution =-=[11, 21, 23]-=-: the computer explores the space of possible parameter settings, and the user acts as an objective-function oracle, interactively selecting computer-suggested alternatives for further exploration. A ... |

1 |
Synthetic lighting for photography. URL http://www.sgi.com/gra ca/synth/index.html
- Haeberli
- 1992
(Show Context)
Citation Context ... turn, these parameters determine the cardinality ofI:jIj = Ph j=1 wj = 584. This particular interface provides additional application-speci c functionality that exploits the additive nature of light =-=[6]-=-. Images can be dragged to the palette, where light intensity and temperature can be varied interactively. Multiple images are composited to form a full-size image in the lower left. 2.4 Results The D... |