harvard panel choices we make
Post on 05-Apr-2018
219 Views
Preview:
TRANSCRIPT
-
7/31/2019 Harvard Panel Choices We Make
1/38
-
7/31/2019 Harvard Panel Choices We Make
2/38
MITComputerScience &ArtificialIntelligenceLaboratory
Introduction
My background
Math/CS
I have read much perception litterature
Amateur photographer
No artistic training, terrible draftsman Two example projects
Photography tonal management
Line drawing from 3D models
-
7/31/2019 Harvard Panel Choices We Make
3/38
MITComputerScience &ArtificialIntelligenceLaboratory
Introduction
I don't build tools, I am an academic researcher, my
deliverables are articles, not software But it's not a complete excuse to write about useless tools
Two types of "consumers/users"
Computer scientists who implement/extend my techniques*Do they understand choices I made, choices they can make?
En-users who use these tools
*Are my choices relevant to them?
-
7/31/2019 Harvard Panel Choices We Make
4/38MITComputerScience &ArtificialIntelligenceLaboratory
Tonal management
Over and under-exposure is the largest
cause of bad photographs Here's a choice I make!
Both for professional and consumers
-
7/31/2019 Harvard Panel Choices We Make
5/38MITComputerScience &ArtificialIntelligenceLaboratory
High Dynamic Range
Real-world contrast is high
Display contrast is low
10-6 106
10-6 106
Real world
Picture
Low contrast
High dynamic range
-
7/31/2019 Harvard Panel Choices We Make
6/38MITComputerScience &ArtificialIntelligenceLaboratory
Our approach
Non-linear two-scale decomposition
Reduce contrast of large scale; preserve local detail
OutputLarge-scale
Detail
Color
-
7/31/2019 Harvard Panel Choices We Make
7/38MITComputerScience &ArtificialIntelligenceLaboratory
Live demo
http://localhost/var/www/apps/conversion/current/tmp/scratch6995/Job/distrib/bila.exe -
7/31/2019 Harvard Panel Choices We Make
8/38MITComputerScience &ArtificialIntelligenceLaboratory
Choices I made
Decomposition: my contribution
Compute in log space: I have good reasons How to separate intensity/color - incidental
How to reduce large-scale layerincidental
Parameters I expose Default parametersmatter of taste
Maybe other choices I don't realize I madeOutputLarge-scale
Detail
Color
-
7/31/2019 Harvard Panel Choices We Make
9/38
A bl h
-
7/31/2019 Harvard Panel Choices We Make
10/38
A programmable approach toLine DrawingWith Stephane Grabli, Emmanuel Turquin &Franois Sillion
-
7/31/2019 Harvard Panel Choices We Make
11/38
MITComputerScience &ArtificialIntelligenceLaboratory
Motivation: Style vs. Technique
Non-Photorealistic Rendering
Imitate traditional media
Each paper focuses on one particular style, which isusually hardcode with a only few available
parameters Stylistic choices mixed with technical ones
-
7/31/2019 Harvard Panel Choices We Make
12/38
MITComputerScience &ArtificialIntelligenceLaboratory
Goal: Decouple style from technique
First step: pure line drawingSystem for rendering line drawing from 3D scenes
Including a flexible style description tool
-
7/31/2019 Harvard Panel Choices We Make
13/38
MITComputerScience &ArtificialIntelligenceLaboratory
First step: pure line drawingSystem for rendering line drawing from 3D scenes
Including a flexible style description tool
Goal: Decouple style from technique
-
7/31/2019 Harvard Panel Choices We Make
14/38
MITComputerScience &ArtificialIntelligenceLaboratory
Goal: Decouple style from technique
First step: pure line drawingSystem for rendering line drawing from 3D scenes
Including a flexible style description tool
Ensuring model independence
-
7/31/2019 Harvard Panel Choices We Make
15/38
MITComputerScience &ArtificialIntelligenceLaboratory
First step: pure line drawingSystem for rendering line drawing from 3D scenes
Including a flexible style description tool
Ensuring model independence
Goal: Decouple style from technique
-
7/31/2019 Harvard Panel Choices We Make
16/38
MITComputerScience &ArtificialIntelligenceLaboratory
Style in line drawing
ITEDO www.itedo.com
Occlusion and nature thickness
-
7/31/2019 Harvard Panel Choices We Make
17/38
MITComputerScience &ArtificialIntelligenceLaboratory
Style in line drawing
Herdman
Depth discontinuity thickness
-
7/31/2019 Harvard Panel Choices We Make
18/38
MITComputerScience &ArtificialIntelligenceLaboratory
Hypothesis
Drawing can be described as a process
Stylistic decisions (line thickness, omission) are
related to scene and image information
These decision strategies can be embedded inprocedures
-
7/31/2019 Harvard Panel Choices We Make
19/38
MITComputerScience &ArtificialIntelligenceLaboratory
Some relevant information
Geometry (2D, 3D coordinates, normals, )
Differential geometry (2D, 3D curvatures, )
Line adjacency
Line nature (silhouette, crease, contour, )
Occluding information(visibility, occluder, depth discontinuity, )
Material
Object id
Drawing density
-
7/31/2019 Harvard Panel Choices We Make
20/38
MITComputerScience &ArtificialIntelligenceLaboratory
Approach
3D View Map
Style
Drawing
+information
-
7/31/2019 Harvard Panel Choices We Make
21/38
MITComputerScience &ArtificialIntelligenceLaboratory
Style: code
Operators.select(QuantitativeInvisibilityUP1D(0))Operators.bidirectionalChain(ChainSilhouetteIterator())Operators.recursiveSplit( Curvature2DF0D(),
pyParameterUP0D(0.2,0.8),NotUP1D(LengthHigherUP1D(75)), 2)
shaders_list = [ StrokeTextureShader("pencil.jpg", Stroke.DRY_MEDIUM, 1),ConstantColorShader(0,0,0,1),ConstantThicknessShader(2.0),
pyGuidingLineShader(),
pyBackboneStretcherShader(0.2)]Operators.create(TrueUP1D(),shaders_list)
classpyGuidingLineShader(StrokeShader):def shade(self, stroke):
it = stroke.strokeVerticesBegin()itlast = stroke.strokeVerticesEnd()itlast.decrement()t = itlast.getObject().getPoint() - it.getObject().getPoint()itmiddle = StrokeVertexIterator(it)
while(itmiddle.getObject().u()
-
7/31/2019 Harvard Panel Choices We Make
22/38
MITComputerScience &ArtificialIntelligenceLaboratory
Style operators
shade
select chain
split
-
7/31/2019 Harvard Panel Choices We Make
23/38
MITComputerScience &ArtificialIntelligenceLaboratory
Style operators
shade
select chain
split
-
7/31/2019 Harvard Panel Choices We Make
24/38
MITComputerScience &ArtificialIntelligenceLaboratory
ShadingThickness Geometry
Color Information dependent
Plain strokes
-
7/31/2019 Harvard Panel Choices We Make
25/38
MITComputerScience &ArtificialIntelligenceLaboratory
Shading
Depth discontinuity thickness
St l t
-
7/31/2019 Harvard Panel Choices We Make
26/38
MITComputerScience &ArtificialIntelligenceLaboratory
Style operators
shade
select chain
split
-
7/31/2019 Harvard Panel Choices We Make
27/38
MITComputerScience &ArtificialIntelligenceLaboratory
Splitting
Split at points of highest 2D curvature
-
7/31/2019 Harvard Panel Choices We Make
28/38
MITComputerScience &ArtificialIntelligenceLaboratory
Splitting
Split at points of highest 2D curvature
Results
-
7/31/2019 Harvard Panel Choices We Make
29/38
MITComputerScience &ArtificialIntelligenceLaboratory
Results
Oriental style
-
7/31/2019 Harvard Panel Choices We Make
30/38
MITComputerScience &ArtificialIntelligenceLaboratory
Oriental style
-
7/31/2019 Harvard Panel Choices We Make
31/38
MITComputerScience &ArtificialIntelligenceLaboratory
Density-based emphasis
-
7/31/2019 Harvard Panel Choices We Make
32/38
MITComputerScience &ArtificialIntelligenceLaboratory
Technical illustration style
Recap
-
7/31/2019 Harvard Panel Choices We Make
33/38
MITComputerScience &ArtificialIntelligenceLaboratory
Recap
Procedural description for style in line drawing
Line drawing from 3D models Control topology, geometry & attributes of strokes
Choices
Automatic picture generation from 3D model Restrict to pure line drawing
Describe style using procedures
View map, types of lines
Information we provide
Types of operators
Choices We Make
-
7/31/2019 Harvard Panel Choices We Make
34/38
MITComputerScience &ArtificialIntelligenceLaboratory
Choices We Make
Model
Algorithms
Parameters
User Interface
Problems we choose
Evaluation criteria
In articles, we must explain the respectiveimportance of choices
It's the question stupid!
-
7/31/2019 Harvard Panel Choices We Make
35/38
MITComputerScience &ArtificialIntelligenceLaboratory
It s the question, stupid!
The important is not the answer to a choice,
it's the choice of the question
And even more importantly, the implicit choices we
make without asking the question
-
7/31/2019 Harvard Panel Choices We Make
36/38
MITComputerScience &ArtificialIntelligenceLaboratory
Other choices
-
7/31/2019 Harvard Panel Choices We Make
37/38
MITComputerScience &ArtificialIntelligenceLaboratory
Other choices
Do we target pro of casual users?
How automatic should things be?
Bad choice consequences
-
7/31/2019 Harvard Panel Choices We Make
38/38
MITComputerScience &ArtificialIntelligenceLaboratory
Bad choice consequences
No uses our technique: we don't address any
problem, or give the wrong solution People are frustrated by our technique
We make something too easy, becomes uniform
Ethical problems
top related