1 / 42

Advanced Image Processing

Advanced Image Processing. Lecture 1 - Introduction 01/14/10. Image Enhancement. Image Segmentation. Image Restoration. Representation & Description. Image Compression. Recognition & Interpretation. Image Coding. Morphological Image Processing. Wavelet Analysis.

donovan
Télécharger la présentation

Advanced Image Processing

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Advanced Image Processing Lecture 1 - Introduction 01/14/10

  2. Image Enhancement Image Segmentation Image Restoration Representation & Description Image Compression Recognition & Interpretation Image Coding Morphological Image Processing Wavelet Analysis Image processing components High-level Image Processing Machine Vision Image Understanding (or Analysis) Low-level Image Processing

  3. Themes and goals • Optimization and consistency • In-depth understanding • Efficient implementation • Distributed processing • Cultivate research style • Research groups • Classic books • Track publications (journal, conf) • Research note

  4. Course organization • Assignments • Homework (63%) • Mostly with programming • Final project (30%) • Algorithm development and paper submission • Final project presentation (7%) • Programming • C++ or Matlab • The Image Processing Library (IPL) • OpenCV

  5. Image enhancement Point-based vs. Neighbor-based processing Spatial domain vs. Frequency domain Image restoration Distortion model Noise model and removal Blur model and removal Least square Constrained least square Low-level image processing • Image compression • Information theory • Lossless • Lossy • Color image processing • Color fundamentals • Pseudo-color generation • Full-color image processing • Morphological operators • Binary morphology • Application • Edge linking • Noise removal • Wavelet analysis • Theory • Application • Image denoising • Image compression

  6. Image enhancement • Point-based • Log transform • Power-law transform • Contrast stretching • Thresholding • Histogram equalization • Image averaging • Mask-based T[f(x,y)] Point processing T[N{f(x,y)}] Mask processing f(x,y) f(x,y) g(x,y) g(x,y)

  7. Log transformation (Dynamic range compression)

  8. Power-Law transformation s = crg

  9. Gamma correction

  10. Contrast stretching

  11. Gray-level slicing Highlighting an intensity range

  12. Histogram equalization

  13. HE – Example 2

  14. HE – Derivation T(r) is single-valued and monotonically increasing within range of r T(r) has the same range as r [0, 1]

  15. Histogram equalization • Transformation function • pr(w) is the probability density function (pdf) • The transformation function is the cumulative distribution function (CDF) • To make the pdf of the transformed image uniform, i.e. to make the histogram of the transformed image uniform

  16. r hist(r) r s 70 10 0 1 2 3 10 80 95 0 1 3 5 70 15 5 10 80 95 15 10 2 100 HE – Discrete case s hist(s) 100

  17. -1 1 -1 1 -1 1 1/9 x -1 1 8 1 1 -1 1/9 x -1 1 1 -1 -1 1 Image enhancement • Mask-based • Lowpass vs. highpass • Linear vs. nonlinear • Lowpass (smoothing) • Average • Weighted average (Gaussian) • Median • Highpass (sharpening) • Unsharp masking (UM) • 1st derivative • 2nd derivative • Point-based • Log transform • Power-law transform • Contrast stretching • Thresholding • Histogram equalization • Image averaging T[N{f(x,y)}] Mask processing T[f(x,y)] Point processing f(x,y) f(x,y) g(x,y) g(x,y)

  18. Unsharp masking and high-boost filters • Unsharp masking • To generate the mask: Subtract a blurred version of the image from itself • Add the mask to the original • Highboost: • k>1 • Application: input image is very dark gmask(x,y) = f(x,y) - f(x,y) g(x,y) = f(x,y) + k*gmask(x,y)

  19. Derivative filters

  20. Filters - 1st derivative • Roberts filter • Prewitt filter • Sobel filter 1 0 0 1 0 -1 -1 0 -1 -1 -1 -1 0 1 0 0 0 -1 0 1 1 1 1 -1 0 1 -1 -2 -1 -1 0 1 0 0 0 -2 0 2 1 2 1 -1 0 1

  21. The Laplacian - Masks 0 1 0 1 1 1 To recover the image: 1 -4 1 1 -8 1 0 1 0 1 1 1 0 -1 0 -1 -1 -1 -1 4 -1 -1 8 1 0 -1 0 -1 -1 -1

  22. Laplacian|Prewitt|Sobel|Roberts Marr-Hildreth|Canny Perfect image Corrupted image

  23. Laplacian|Prewitt|Sobel|Roberts Same with threshold Marr-Hildreth|Canny|Angiogram

  24. Image enhancement • Spatial-based • Frequency-based • Why? • Computational efficiency • More intrinsic information • Lowpass vs. highpass • Notch filters

  25. 2-D Fourier transform • CFT • DFT

  26. Practical issues – Implement convolution in frequency domain • In spatial domain • In frequency domain • f*g  F(f)G(g) • Phase? Mag? • How to pad?

  27. Difference image from convolution in the spatial domain Convolution in the frequency domain No padding With padding Conv. spatially

  28. Lowpass filtering • Ideal filter • D(u, v): distance from point (u, v) to the origin • cutoff frequency (D0) • nonphysical • radially symmetric about the origin • Butterworth filter • Gaussian lowpass filter

  29. Power ratio 99.9699.6599.0497.84

  30. Highpass filter • Ideal filter • Butterworth filter • Gaussian highpass filter

  31. Example

  32. Optimal notch filters

  33. Optimal notch - Derivation

More Related