Information-Geometric Lenses for Multiple Foci + Contexts Interfaces Richard Nock Frank Nielsen CEREGMIA – UAG Sony CS Labs France [email protected] Japan [email protected]
Information-Geometric Lenses for Multiple Foci + Contexts Interfaces
Dec 07, 2014
Presented at ACM SIGGRAPH ASIA 2013 along with videos
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Information-Geometric Lenses for
Multiple Foci + Contexts Interfaces
Richard Nock Frank Nielsen
CEREGMIA – UAG Sony CS Labs
France
Japan
Outline
F+C interfaces
Definition
Example of desirable property
InfoGeoLenses
Definition & principles
Properties
Applications
2D InfoGeoLenses
3D InfoGeoLenses
F+C interfaces
F+C Interface > Definition
Integrate a focus (magnified) region and its context in a seamless
display
Focus region
Context
(From Cockburn et al., 2007)
Example > Reliable perception of shapes
The Euclidean shape of a perceived disk in the F+C interface
?
Focus
Example > Reliable perception of shapes
The Euclidean shape of a perceived disk in the F+C interface
Sarkar-Brown ‘92
SB(0) disk
Example > Reliable perception of shapes
The Euclidean shape of a perceived disk in the F+C interface
Sarkar-Brown ‘92
SB(2) ~disk
Example > Reliable perception of shapes
The Euclidean shape of a perceived disk in the F+C interface
Sarkar-Brown ‘92
SB(5) disk
Example > Reliable perception of shapes
The Euclidean shape of a perceived disk in the F+C interface
InfoGeoLens
LOG ~disk
Example > Reliable perception of shapes
The Euclidean shape of a perceived disk in the F+C interface
InfoGeoLens
MAT ~disk
InfoGeoLenses
InfoGeoLenses > Definition & principles
Map each coordinate (x, y, z, …) through a “connection map”:
function
continuous,
differentiable
strictly increasing,
symmetric wrt .
Different functions yield different InfoGeoLenses
Logistic, LOG
Matsushita, MAT
Modified Sarkar Brown, MSB
(Prop: MSB = 1st order approximation of LOG)
InfoGeoLenses > Definition & principles
InfoGeoLens
MSB
InfoGeoLenses > Definition & principles
Examples of InfoGeoLenses
MSB =
Modified Sarkar-Brown
InfoGeoLenses > Properties
Region near the focus is approximately Euclidean
InfoGeoLenses > Properties
Simple and almost lossless reconstruction of the original image
through
MSB
InfoGeoLenses > Properties
Simple and almost lossless reconstruction of the original image
through
Original
image
reconstructed
InfoGeoLenses > Properties
Simple and almost lossless reconstruction of the original image
through
MSB
InfoGeoLenses > Properties
Simple and almost lossless reconstruction of the original image
through
Original
image
reconstructed
InfoGeoLenses > Properties
Simple and almost lossless reconstruction of the original image
through
MSB
InfoGeoLenses > Properties
Simple and almost lossless reconstruction of the original image
through
Original
image
reconstructed
InfoGeoLenses > Properties
Simple construction of multiple foci+context interfaces
LOG
Voronoi
diagram
InfoGeoLenses > Properties
Simple construction of multiple foci+context interfaces
MAT
Voronoi
diagram
Applications > 2D InfoGeoLens
InfoGeoLenses > Demo 2D F+C interface
Virtual library (1k books) embedded in the classical
Beltrami-Klein disk, and then LOG, then MAT interface
Red dots = books; hexagons = language (English/French)
InfoGeoLenses > Demo 2D F+C interface
Applications > 3D InfoGeoLens
InfoGeoLenses > Demo 3D F+C interface
Virtual library (10k books) embedded in the LOG cube
Rich interaction (zoom, rotations & translations can be inside/outside)
3D representation dynamically learned from users behavior (see
Nock, Nielsen & Briys, ACM/IEEE JCDL’2013); colors = Dewey
Subjects
InfoGeoLenses > Demo 3D F+C interface
Conclusion
InfoGeoLenses inspired by Information Geometry
Definition
Example of desirable property
InfoGeoLenses useful for blending with other interfaces
Definition & principles
Properties
Thank you > Questions ?
Related Documents
