1. Time-Frequency Analysis and Wavelet Transform Oral Presentation Advisor: 丁建均 and All Class Members Student: 李境嚴 ID: D00945001

2. What’s Today?

3. 3 (1) Finding Instantaneous Frequency (2) Signal Decomposition (3) Filter Design (4) Sampling Theory (5) Modulation and Multiplexing (6) Electromagnetic Wave Propagation (7) Optics (8) Radar System Analysis (9) Random Process Analysis (10) Music Signal Analysis (11) Acoustics (12) Biomedical Engineering (13) Spread Spectrum Analysis (14) System Modeling (15) Image Processing (16) Economic Data Analysis (17) Signal Representation (18) Data Compression (19) Seismology (20) Geology XIII. Applications of Time–Frequency Analysis Biomedical Engineering Image Processing Wavelet Transform Kernel (Windows) Laws Texture

4.  Study of Classification of Lung Tumor Based on CT/PET Images  Technique of studying image ( gray level)  Training skill of machine learning What’s Today?

5.  Gray level studying  DSP, Kernel( window)  Resolution of image  4000*3000, 1024*768, 640*480, 320*240  How about in Biomedical Image? Why Image Processing?

6.  The Biomedical Image Today  CT:  512*512  PET: 128*128 Why Image Processing?

7. Brain v.s. Lung Tumors

8.  Introduction and Back ground  Technique  Experiments  Discussion and Conclusion Outline

9.  Introduction and Back ground  Technique  Experiments  Discussion and Conclusion Introduction

10.  Lung Tumor  High Death Ratio  Nerve-less Introduction

11. Image Load Pre- processing Co- Registration ROI Feature Extraction Classification Co- Registration Feature Extraction Down / Up sampling ; Wavelet Transform Wavelet ; Laws Texture ; Other Methods

12.  Wavelet Transform: Introduction--Wavelet Transform J. J. Ding, 09 月 15 日上課資料, P 43

13. Introduction--Wavelet Transform Ivan W. Selesnick, Wavelet Transforms, 2007

14. Introduction--Wavelet Transform

15. Introduction Ivan W. Selesnick, Wavelet Transforms, 2007

16. Introduction--Wavelet Transform Ivan W. Selesnick, Wavelet Transforms, 2007

17. Introduction--Wavelet Transform Ivan W. Selesnick, Wavelet Transforms, 2007

18.  Wavelet Transform:  Improvement???  Haar !! Introduction--Wavelet Transform

19.  Haar Transform: Introduction--Wavelet Transform

20. Wavelet Transform Haar Transform

21.  Wavelet Transform: Introduction--Wavelet Transform J. J. Ding, 09 月 15 日上課資料, P 46

22. Introduction--Wavelet Transform

23.  Laws features  The texture energy measures developed by Kenneth Ivan Laws at the University of Southern California have been used for many diverse applications. These measures are computed by first applying small convolution kernels to a digital image, and then performing a nonlinear windowing operation. Introduction—Laws Texture http://www.ccs3.lanl.gov/~kelly/ZTRANSITION/notebook/laws.shtml

24.  Laws features  3 element kernel  5 element kernel  High order kernel Introduction—Laws Texture M.T. Suzuki, Y. Yaginuma, H. Kodama, A Texture Energy Measurement Technique for 3D Volumetric Data, 2009 IEEE International Conference on Systems http://www.ccs3.lanl.gov/~kelly/ZTRANSITION/notebook/laws.shtml

25.  Laws features  3 element kernel  Level: [1 2 1];  Edge:[-1 0 1];  Spot:[-1 2 -1]; Introduction—Laws Texture M.T. Suzuki, Y. Yaginuma, H. Kodama, A Texture Energy Measurement Technique for 3D Volumetric Data, 2009 IEEE International Conference on Systems http://www.ccs3.lanl.gov/~kelly/ZTRANSITION/notebook/laws.shtml

26.  Laws features Introduction—Laws Texture

27.  Laws features  5 element kernel  L5 = [1, 4, 6, 4, 1];  E5 = [−1,−2, 0, 2, 1];  S5 = [−1, 0, 2, 0,−1];  R5 = [1,−4, 6,−4, 1];% ripple  W5 = [−1, 2, 0,−2, 1];% wave Introduction—Laws Texture M.T. Suzuki, Y. Yaginuma, H. Kodama, A Texture Energy Measurement Technique for 3D Volumetric Data, 2009 IEEE International Conference on Systems http://www.ccs3.lanl.gov/~kelly/ZTRANSITION/notebook/laws.shtml

28.  Laws features Introduction—Laws Texture

29.  Laws features  Image processing --- 2D case L5L5 L5E5 L5S5 L5R5 L5W5 E5L5 E5E5 E5S5 E5R5 E5W5 S5L5 S5E5 S5S5 S5R5 S5W5 R5L5 R5E5 R5S5 R5R5 R5W5 W5L5 W5E5 W5S5 W5R5 W5W5 Introduction—Laws Texture M.T. Suzuki, Y. Yaginuma, H. Kodama, A Texture Energy Measurement Technique for 3D Volumetric Data, 2009 IEEE International Conference on Systems http://www.ccs3.lanl.gov/~kelly/ZTRANSITION/notebook/laws.shtml

30.  Laws features Introduction—Laws Texture

31. L5L5 E5E5 S5S5 R5R5 W5W5

32.  CT - computed tomography  PET - Positron emission tomography Introduction-- Background

33.  CT - Computed Tomography  Digital geometry processing is used to generate a three-dimensional image of the inside of an object from a large series of two-dimensional X-ray images taken around a single axis of rotation.  http://translate.google.com/translate?hl=zh-TW&langpair=en|zh- TW&u=http://en.wikipedia.org/wiki/X-ray_computed_tomography Introduction-- Background

34.  PET - Positron Emission Tomography  A nuclear medicine imaging technique that produces a three-dimensional image or picture of functional processes in the body. The system detects pairs of gamma rays emitted indirectly by a positron-emitting radionuclide (tracer), which is introduced into the body on a biologically active molecule. Three-dimensional images of tracer concentration within the body are then constructed by computer analysis. In modern scanners, three dimensional imaging is often accomplished with the aid of a CT X-ray scan performed on the patient during the same session, in the same machine.  If the biologically active molecule chosen for PET is FDG, an analogue of glucose, the concentrations of tracer imaged then give tissue metabolic activity, in terms of regional glucose uptake. Although use of this tracer results in the most common type of PET scan, other tracer molecules are used in PET to image the tissue concentration of many other types of Introduction-- Background http://en.wikipedia.org/wiki/Positron_Emission_Tomography

35.  PET - Positron emission tomography  FDG ( Fludeoxyglucose) :  氟代脱氧葡萄糖 Introduction-- Background http://en.wikipedia.org/wiki/Positron_Emission_Tomography

36. Background CT V.S. PET

37.  Introduction and Back ground  Technique  Experiments  Discussion and Conclusion Technique

38.  Feature Extracting – 1 (on CT)  Down sampling (for co-registry)  Overlap CT/PET( Down/Up Sampling)  Feature Extracting – 2 (on PET)  Machine Learning Technique

39. Background CT V.S. PET

40.  Feature Extracting – 1 (on CT)  Volume  Rectangular Fit  Histogram features  Laws features  Wavelet ： Technique – Feature Extracting – 1 (on CT)

41. Technique – Feature Extracting – 1 (Wavelet) 2D Case

42. Technique – Feature Extracting – 1 (Wavelet) 3D Case

43. Technique – Feature Extracting – 1 (Laws Texture) 2D Case

44. Technique – Feature Extracting – 1 (Laws Texture) 3D Case

45.  Down sampling (for co-registry) Technique – Down sampling (for co-registry) Raw Image Low Pass (Average) High Pass 1 (X direction) High Pass 2 (Y direction) High Pass 3 (Corner)

46.  Down sampling (for co-registry) Technique – Down sampling (for co-registry) Raw Image Low Pass (Average) High Pass 1 (X direction) High Pass 2 (Y direction) High Pass 3 (Corner) Down-samples Image

47.  Feature Extracting – 2 (on PET)  SUV  Leveled SUV  Largest Region’s SUV  Other probability features Technique – Feature Extracting – 2 (on PET)

48.  Feature Extracting – 2 (on PET) Technique – Feature Extracting – 2 (on PET) PAWITRA MASA-AH, SOMPHOB SOONGSATHITANON, A novel Standardized Uptake Value (SUV) calculation of PET DICOM files using MATLAB, NEW ASPECTS OF APPLIED INFORMATICS, BIOMEDICAL ELECTRONICS & COMMUNICATIONS

49.  Feature Extracting – 2 (on PET) Technique – Feature Extracting – 2 (on PET) Tumor Level 1 Sub SUV Level 2 Sub SUV Level 3 Sub SUV Level 4 Sub SUV Level 5 Sub SUV Feature

50.  Machine Learning  Logistic  Neural Network  SVM (Support Vector Machine)  J48 Technique – Machine Learning

51.  Introduction  Background  Technique  Experiments  Discussion and Conclusion Experiments

52.  Sorry, they are now in America Experiments

53.  Introduction  Background  Technique  Experiments  Discussion and Conclusion Discussion and Conclusion

54.  Discussion:  Relation between Image Processing, DSP, and TWD  Kernel of Image Processing  Development of Each kernel Discussion and Conclusion

55.  Relation between Image Processing, DSP, and TWD  TWD:  Analyzing signal with mathematically way, either enhancement of complexity of equation and reducing the amount of computation.  DSP:  Dealing the signal with discrete time work.  DIP:  Take advantage of these two to give us more probabilities on studying images. Discussion and Conclusion

56.  Kernel of Image Processing  Similar to the window function on short time signal analysis  Either Gaussian filter (low pass filtering, averaging) and edge detection (high pass filtering) are applied to turn into features Discussion and Conclusion

57.  Development of Each kernel  Low pass filter  High pass filter Discussion and Conclusion

58.  Development of Each kernel  Low pass filter  Down sample ( average)  [1 1]  Laws texture (level)  [1 2 1], [1 4 6 4 1]  Gaussian blur (normal distribution)  [1 8 12 16 12 8 1] Discussion and Conclusion

59.  Development of Each kernel  High pass filter  Down sample ( change)  [1 -1]  Laws texture (edge, ripple)  [-1 -2 0 2 1], [1,−4, 6,−4, 1]  Gaussian Laplace Filter  Subtract by two Gaussian filter with same mean, different STD. Discussion and Conclusion

60.  Development of Each kernel  High pass filter  Down sample ( change)  [1 -1]  Laws texture (edge, ripple)  [-1 -2 0 2 1], [1,−4, 6,−4, 1]  Gaussian Laplace Filter  Subtract by two Gaussian filter with same mean, different STD. Discussion and Conclusion

61.  Development of Each kernel  High pass filter Discussion and Conclusion

62.  Development of Each kernel  High pass filter Discussion and Conclusion

63.  Conclusion:  Image processing is right an example which implement DSP and TWD.  Texture Feature give doctors more clues for diagnosing  More kinds of kernel provide more feature for machine learning. Discussion and Conclusion