Simulation of Fibrous Scaffold Optimal Distribution by Genetic Algorithm Presentation : D. Semnani ICSIP 2009, Amsterdam Isfahan University of Technology.

Slides:



Advertisements
Similar presentations
G. Lucas and Xin Zhao University of Huddersfield, UK
Advertisements

Regional Processing Convolutional filters. Smoothing  Convolution can be used to achieve a variety of effects depending on the kernel.  Smoothing, or.
Acoustic design by simulated annealing algorithm
CDS 301 Fall, 2009 Image Visualization Chap. 9 November 5, 2009 Jie Zhang Copyright ©
Image Segmentation Image segmentation (segmentace obrazu) –division or separation of the image into segments (connected regions) of similar properties.
Automatic measurement of pores and porosity in pork ham and their correlations with processing time, water content and texture JAVIER MERÁS FERNÁNDEZ MSc.
Optimal Design Laboratory | University of Michigan, Ann Arbor 2011 Design Preference Elicitation Using Efficient Global Optimization Yi Ren Panos Y. Papalambros.
Pore Detection in Small Diameter Bores The University of Michigan, Ann Arbor NSF Engineering Research Center for Reconfigurable Manufacturing Systems.
Light and Reflection Level 1 Physics. Facts about Light It is a form of Electromagnetic Energy It is a part of the Electromagnetic Spectrum and the only.
Dendrochronology Sebastian Hegenbart Joachim Kerschbaumer
Institute of Intelligent Power Electronics – IPE Page1 Introduction to Basics of Genetic Algorithms Docent Xiao-Zhi Gao Department of Electrical Engineering.
1Ellen L. Walker Edges Humans easily understand “line drawings” as pictures.
Edge Detection CSE P 576 Larry Zitnick
Lecture 4 Edge Detection
COMP305. Part II. Genetic Algorithms. Genetic Algorithms.
Segmentation Divide the image into segments. Each segment:
Edge Detection Today’s reading Forsyth, chapters 8, 15.1
COMP305. Part II. Genetic Algorithms. Genetic Algorithms.
Edge Detection Today’s readings Cipolla and Gee –supplemental: Forsyth, chapter 9Forsyth Watt, From Sandlot ScienceSandlot Science.
© 2010 Cengage Learning Engineering. All Rights Reserved.
Chapter 6: Transform and Conquer Genetic Algorithms The Design and Analysis of Algorithms.
Measurement of Nonwoven Surface Roughness With Machine Vision Method Presentation : D. Semnani ICSIP 2009, Amsterdam Isfahan University of Technology.
Spatial-based Enhancements Lecture 3 prepared by R. Lathrop 10/99 updated 10/03 ERDAS Field Guide 6th Ed. Ch 5: ;
National Center for Supercomputing Applications University of Illinois at Urbana-Champaign Image Features Kenton McHenry, Ph.D. Research Scientist.
Genetic Algorithm.
SOFT COMPUTING (Optimization Techniques using GA) Dr. N.Uma Maheswari Professor/CSE PSNA CET.
Hierarchical Distributed Genetic Algorithm for Image Segmentation Hanchuan Peng, Fuhui Long*, Zheru Chi, and Wanshi Siu {fhlong, phc,
Intro. ANN & Fuzzy Systems Lecture 36 GENETIC ALGORITHM (1)
Boltzmann Machine (BM) (§6.4) Hopfield model + hidden nodes + simulated annealing BM Architecture –a set of visible nodes: nodes can be accessed from outside.
Digital Image Processing CCS331 Relationships of Pixel 1.
Edges. Edge detection schemes can be grouped in three classes: –Gradient operators: Robert, Sobel, Prewitt, and Laplacian (3x3 and 5x5 masks) –Surface.
Applying Genetic Algorithm to the Knapsack Problem Qi Su ECE 539 Spring 2001 Course Project.
Optimal Placement of Wind Turbines Using Genetic Algorithms
Chapter 10 Image Segmentation.
1 A New Method for Composite System Annualized Reliability Indices Based on Genetic Algorithms Nader Samaan, Student,IEEE Dr. C. Singh, Fellow, IEEE Department.
A New Evolutionary Approach for the Optimal Communication Spanning Tree Problem Sang-Moon Soak Speaker: 洪嘉涓、陳麗徽、李振宇、黃怡靜.
Mathematical Morphology Mathematical morphology (matematická morfologie) –A special image analysis discipline based on morphological transformations of.
Many slides from Steve Seitz and Larry Zitnick
DIVERSITY PRESERVING EVOLUTIONARY MULTI-OBJECTIVE SEARCH Brian Piper1, Hana Chmielewski2, Ranji Ranjithan1,2 1Operations Research 2Civil Engineering.
Edge Assembly Crossover
Genetic Algorithms What is a GA Terms and definitions Basic algorithm.
Genetic Algorithms Abhishek Sharma Piyush Gupta Department of Instrumentation & Control.
Chap11 Basic characteristics parameter of yarn 11.1 Fineness and unevenness of yarn 11.1 Fineness and unevenness of yarn 11.2 Twist and twist shrinkage.
The law of reflection: The law of refraction: Image formation
Image Processing in Textile Dariush SemnaniMorteza Vadood Isfahan University of Technology.
Authors: Soamsiri Chantaraskul, Klaus Moessner Source: IET Commun., Vol.4, No.5, 2010, pp Presenter: Ya-Ping Hu Date: 2011/12/23 Implementation.
Genetic Algorithm Dr. Md. Al-amin Bhuiyan Professor, Dept. of CSE Jahangirnagar University.
2/5/16Oregon State University PH 212, Class #151 Snell’s Law This change in speed when light enters a new medium means that its wavefronts will bend, as.
►Search and optimization method that mimics the natural selection ►Terms to define ٭ Chromosome – a set of numbers representing one possible solution ٭
Artificial Intelligence By Mr. Ejaz CIIT Sahiwal Evolutionary Computation.
Processing Images and Video for An Impressionist Effect Automatic production of “painterly” animations from video clips. Extending existing algorithms.
Genetic Algorithms An Evolutionary Approach to Problem Solving.
Presented By: Farid, Alidoust Vahid, Akbari 18 th May IAUT University – Faculty.
Intelligent Database Systems Lab 國立雲林科技大學 National Yunlin University of Science and Technology 1 Intelligent Exploration for Genetic Algorithms Using Self-Organizing.
Digital Image Processing (Digitaalinen kuvankäsittely) Exercise 5
Refraction and Lenses.
Chapter 10 Image Segmentation
In Search of the Optimal Set of Indicators when Classifying Histopathological Images Catalin Stoean University of Craiova, Romania
CS621: Artificial Intelligence
Modified Crossover Operator Approach for Evolutionary Optimization
Sunny Ri Li, Nasser Ashgriz
ECE 692 – Advanced Topics in Computer Vision
Computer Vision Lecture 16: Texture II
Aim of the project Take your image Submit it to the search engine
Thin lens formula The thin lens formula relates the focal length of a lens to the object and image distances. If two of these properties are known,
Evolving Logical-Linear Edge Detector with Evolutionary Algorithms
Boltzmann Machine (BM) (§6.4)
Edge Detection Today’s readings Cipolla and Gee Watt,
Introduction to Genetic Algorithm and Some Experience Sharing
Introduction to Artificial Intelligence Lecture 22: Computer Vision II
Presentation transcript:

Simulation of Fibrous Scaffold Optimal Distribution by Genetic Algorithm Presentation : D. Semnani ICSIP 2009, Amsterdam Isfahan University of Technology

2 Artificial Intelligence in Textile Engineering ICSIP 2009, Amsterdam Optimization in Textiles Process, Material and Machinery Classification of Products Measuring Uniformity of Fibrous Structures Determination of Woven And Nonwoven Fabrics Characteristics Prediction of Tissue Quality

3 Application of image processing to determine the properties of the non- woven and nano-fibers Lighting Light source with low wave length, laser and light emitting diodes Magnification Preparation Value of the each pixel based on the adjacent values

4 SKELETONIZING OR THINNING Replacing each object with a narrow line (thickness: 1 pixel) Morphological or Pruning

5 FIBER ORIENTATION DISTRIBUTION (FOD) Orientation distribution function α: the angle between fiber and horizontal axis Fibers were made from direct and short lines Creating artificial images and testing algorithms Comparing obtained results

6 Direct tracking Furrior transform Hough transform Flow Field Analys Orientation measuring methods

7 DIRECT TRACKING SEARCH Using Morphological or Pruning methods Every pixel has 8 adjacent pixels

8 DIRECT TRACKING SEARCH It is assumed that the fibers are one pixel thick and have not severe disruptions or kinks or bends within one pixel distance.

9 FOURIOR TRANSFORM An image web was formed from light cycles (dark to white and vice versa) u : frequency in X axis v : frequency in Y axis

10 FURRIOR TRANSFORM Power Spectrum Function If fiber are orientated in a special direction so frequency in same direction is low and in perpendicular direction is high.

11 FURRIOR TRANSFORM Evaluating image by special radius and loop thickness

12 FURRIOR TRANSFORM If the image is not periodic, then discontinuation points appear in transformed image.

13 FLOW FIELD ANALISYS The edges of image present the field orientations stages 1.Morphological operation 2.Calculating gradient vector for all points 3.Dividing image to the small images 4.Determining the mean orientation of fields in each small image 5.Calculating the final image orientation by using mean orientations of small images 13

14 GAUSSIAN FILTER Replacing each point by regarding adjacent points H and W : size of kernal matrice

15 GRADIENT Sobel matrice Gy Gy Gx Gx x7x7 x4x4 x1x1 i-1 x8x8 x5x5 x2x2 i x9x9 x6x6 x3x3 i+1 j+1jj-1

16 FLOW FIELD ANALISYS

17 HOUGH TRANSFORM

18 HOUGH TRANSFORM

19 COMPARING METHODS Direct Tracking is the best method for on-line controlling but it has low speed process because of loops in its algorithm. Flow Field Analisys evaluate the STD lower than the other methods and can be used in on-line controlling. Furrior Transform is the best choice to non-on-line controlling. The results of Hough and Furrior Transform is so close. Direct Tracking Furrior Transform Flow Field Analisys Accuracy ranking

20 ORIENTATION IN REAL WEB The best image will be one that represents the entire field as a two dimensional projection.

21 EDGE THRESHOLDING

22 FIBER DIAMETER DISTRIBUTION

23 FIBER DIAMETER DISTRIBUTION

24 Threshold 2Threshold 1Threshold 3 MEASURING THE POROSITY OF VARIOUS SURFACE LAYERS Threshold 1 : Threshold 2 : Threshold 3 :

25 MEASURING THE POROSITY OF VARIOUS SURFACE LAYERS n :Number white points N : Number of all points p : Porosity percentage

26 C ALCULATING THE P OROSITY

27 LAYER UNIFORMITY

28 L AYER U NIFORMITY

29 MEASURING LAYER WEIGHT

30 Our Method ICSIP 2009, Amsterdam Ideal Structure

31 GA ICSIP 2009, Amsterdam Optimizing the model SELECTION : selecting individuals for reproduction. REPRODUCTION: Cross over and Mutation are most common reproduction operators of GA. EVALUATION: the fitness of new chromosome is evaluated. REPLACEMENT: individuals from the old population are removed and replaced by the new ones. The algorithm is stopped when the population converges toward optimal solution e.g. finding minimum of a function

32 GA MOdel ICSIP 2009, Amsterdam The Number of lines in each group was equal. The Chromosomes have defined in a binary from There were two Genes with lengths of 5 and 19 bits The angle drops between

33 Image Processing for Fitness ICSIP 2009, Amsterdam Plotting the structure

34 Real Web ICSIP 2009, Amsterdam Histogram Modification, Thresholding, Converting to binary form and Thinning.

35 Optimal Web ICSIP 2009, Amsterdam

36 ICSIP 2009, Amsterdam Comparison optimal model and real web with real web Find the optimal web Fitness measuring Real web production Image Enhancement Measuring the fitness of real web Comparison optimal model and real web Analysis mechanical properties

37 ICSIP 2009, Amsterdam Breaking load and fitness value VS Sample

38 Conclusion Simulated a non-woven web with optimal distribution, using Genetic Algorithm Relationship between distribution uniformity of a web and its breaking load Validity of such a relation has been investigated by performing the fitness function In another words, the sample which were more uniform, had a higher breaking load. In a further research we will investigate this relationship on a three dimensional structure of a fiber reinforcement composite. ICSIP 2009, Amsterdam