Genetic Algorithm with Knowledge-based Encoding for Interactive Fashion Design

1
Genetic Algorithm with Knowledge-based Encoding
for Interactive Fashion Design
PRICAI-2000

• Hee-Su Kim and Sung-Bae Cho
• Computer Science Department, Yonsei University
• Shinchon-dong, Sudaemoon-ku, Seoul 120-749, Korea
• madoka, sbcho_at_candy.yonsei.ac.kr

2
Agenda

• Motivation
• Backgrounds
• System development
• Knowledge-based encoding
• Experimental results
• Conclusion and future works

3
Motivation
Changes on Consumer Economy
Before the Industrial Revolution Customers
have few choices on buying their clothes
Manufacturer Oriented
After the Industrial Revolution Customers can
make their choices with very large variety
Near Future Customers can order and get clothes
of their favorite design
Consumer Oriented
4
Motivation
Need for Interaction-based System

• Almost all consumers are non-professional at
  design
• To make designers contact all consumers is not
  effective
• Need for the design system that can be used by
  non-professionals

5
Fashion Design
Backgrounds

• Fashion design
• To make a choice within various styles that
  clothes can take
• Three shape part of fashion design
• Silhouette
• Detail
• Trimming

6
Genetic Algorithm
Backgrounds

• t0
• Initialize Population
• Evaluate P(t)
• while not done do
• tt1
• PSelect Parents P(t)
• Recombine P(t)
• Mutate P(t)
• Evaluate P(t)
• PSurvive P, P(t)
• end while

Crossover
Mutation
7
Interactive Genetic Algorithm
Backgrounds
8
Related Works
Backgrounds

• Virtuosi System (Nottingham Trent University,
  1998)
• AutoCAD with ApparelCAD plug-in (Autodesk co.)
• Fashion design aid system for professionals only
• Manual Evolutionary Design Aid System (Nakanishi,
  1996)
• Often produces impractical designs

Apply evolutionary Computation using domain
specific knowledge
Interactive GA
KB Encoding
9
Overview
System Development
10
System Development
3D Modeling Method

• VRML Simply get 3D but too slow
• OpenGL Faster but not easy to implement
• Use GLUT library with OpenGL
• Reduce the burden of programming OpenGL

11
Modeling by 3D Studio MAX
System Development
12
IGA Fashion Design Aid System
System Development
13
Gene Encoding
Knowledge-based Encoding
A
B
C
D
E
F
Total 23 bits
E Skirt and waistline style(9)
F Color(8)
B Color(8)
A Neck and body style(34)


D Color(8)
C Arm and sleeve style(11)
Search space size 34811898 1,880,064

14
Example Design from a Genotype
Knowledge-based Encoding
15
Schema Theorem
Knowledge-based Encoding

• The instances of schema H in particular
  generation t1, m(H, t1), can be expressed in
  terms of m(H, t)
• Schemata with short defining length, low order,
  above-average fitness receive exponentially
  increasing trials in subsequent generations

16
Experimental Environment
Experimental Results

• Subjects
• 10 male and female student, no background on
  fashion design
• Crossover rate 0.5 (1-point crossover)
• Mutation rate 0.05 (Binary mutation)
• 10 generations with elitist preserving
• Request for each subjects
• Find out most cool-looking design with given
  system

17
Convergence Test for Cool-Looking Design
Experimental Results
18
Subjective Test
Experimental Results
Examples of searched design which gives cool
feeling
19
Fitness Changes for each Encoding Method
Experimental Results
20
Relative Satisfaction for each Encoding Method
Experimental Results
21
Example Solution Design and Frequency of Each
Solution Schema
Experimental Results
22
Conclusion and Future Works

• Knowledge-based Encoding in Interactive Genetic
  Algorithm for a Fashion Design Aid System
• Based on Knowledge of fashion design
• Compared with sequential encoding by several
  experiments
• Future Works
• Adding up extra design elements such as textile
  To enlarge the search space
• Clustering To avoid Genetic drift caused by
  small population size
• Direct Manipulation To accelerate convergence
  with relatively short generation