2007Theo Schouten1 Enhancements Techniques for editing an image such that it is more suitable for a specific application than the original image. Spatial.

Slides:



Advertisements
Similar presentations
Linear Filtering – Part I Selim Aksoy Department of Computer Engineering Bilkent University
Advertisements

Spatial Filtering (Chapter 3)
Topic 6 - Image Filtering - I DIGITAL IMAGE PROCESSING Course 3624 Department of Physics and Astronomy Professor Bob Warwick.
Image Processing Lecture 4
CS & CS Multimedia Processing Lecture 2. Intensity Transformation and Spatial Filtering Spring 2009.
Spatial Filtering.
Local Enhancement Histogram processing methods are global processing, in the sense that pixels are modified by a transformation function based on the gray-level.
Image Filtering. Outline Outline Concept of image filter  Focus on spatial image filter Various types of image filter  Smoothing, noise reductions 
Chapter 3 Image Enhancement in the Spatial Domain.
Lecture 6 Sharpening Filters
Frequency Domain Filtering (Chapter 4)
Image Enhancement in the Frequency Domain Part III
Image processing (spatial &frequency domain) Image processing (spatial &frequency domain) College of Science Computer Science Department
© by Yu Hen Hu 1 ECE533 Digital Image Processing Image Enhancement in Frequency Domain.
EE 4780 Image Enhancement. Bahadir K. Gunturk2 Image Enhancement The objective of image enhancement is to process an image so that the result is more.
Digital Image Processing
E.G.M. PetrakisFiltering1 Linear Systems Many image processing (filtering) operations are modeled as a linear system Linear System δ(x,y) h(x,y)
Digital Image Processing In The Name Of God Digital Image Processing Lecture3: Image enhancement M. Ghelich Oghli By: M. Ghelich Oghli
Chapter 3: Image Enhancement in the Spatial Domain
6/9/2015Digital Image Processing1. 2 Example Histogram.
Chapter 4 Image Enhancement in the Frequency Domain.
Machine Vision and Dig. Image Analysis 1 Prof. Heikki Kälviäinen CT50A6100 Lectures 6&7: Image Enhancement Professor Heikki Kälviäinen Machine Vision and.
Digital Image Processing
MSU CSE 803 Stockman Linear Operations Using Masks Masks are patterns used to define the weights used in averaging the neighbors of a pixel to compute.
1 Vladimir Botchko Lecture 4. Image Enhancement Lappeenranta University of Technology (Finland)
Enhancing Images Ch 5:Shapiro, Ch 3:Gonzales. Gray level Mapping Brightness Transform: 1. Position Dependent f(i,j)= g(i,j). e(i,j) g:Clean image e:position.
Lecture 2. Intensity Transformation and Spatial Filtering
MSU CSE 803 Linear Operations Using Masks Masks are patterns used to define the weights used in averaging the neighbors of a pixel to compute some result.
Chapter 4 Image Enhancement in the Frequency Domain.
Chapter 3 Image Enhancement in the Spatial Domain.
Image Filtering. Problem! Noise is a problem, even in images! Gaussian NoiseSalt and Pepper Noise.
ECE 472/572 - Digital Image Processing Lecture 4 - Image Enhancement - Spatial Filter 09/06/11.
Chapter 3 (cont).  In this section several basic concepts are introduced underlying the use of spatial filters for image processing.  Mainly spatial.
Chap2 Image enhancement (Spatial domain)
Introduction to Image Processing Grass Sky Tree ? ? Review.
Presentation Image Filters
Machine Vision ENT 273 Image Filters Hema C.R. Lecture 5.
Spatial Filtering: Basics
Medical Image Analysis Image Enhancement Figures come from the textbook: Medical Image Analysis, by Atam P. Dhawan, IEEE Press, 2003.
Chapter 3 Image Enhancement in the Spatial Domain.
Lecture 03 Area Based Image Processing Lecture 03 Area Based Image Processing Mata kuliah: T Computer Vision Tahun: 2010.
Image Processing is replacing Original Pixels by new Pixels using a Transform rst uvw xyz Origin x y Image f (x, y) e processed = v *e + r *a + s *b +
ENG4BF3 Medical Image Processing Image Enhancement in Frequency Domain.
Chapter 5: Neighborhood Processing
Digital Image Processing (Digitaalinen kuvankäsittely) Exercise 2
Machine Vision ENT 273 Image Filters Hema C.R. Lecture 5.
Spatial Filtering.
Image Subtraction Mask mode radiography h(x,y) is the mask.
Low Pass Filter High Pass Filter Band pass Filter Blurring Sharpening Image Processing Image Filtering in the Frequency Domain.
Mestrado em Ciência de Computadores Mestrado Integrado em Engenharia de Redes e Sistemas Informáticos VC 15/16 – TP7 Spatial Filters Miguel Tavares Coimbra.
İmage enhancement Prepare image for further processing steps for specific applications.
Machine Vision Edge Detection Techniques ENT 273 Lecture 6 Hema C.R.
EE 7730 Image Enhancement. Bahadir K. Gunturk2 Image Enhancement The objective of image enhancement is to process an image so that the result is more.
Image Enhancement by Spatial Domain Filtering
Frequency Domain Filtering. Frequency Domain Methods Spatial Domain Frequency Domain.
Digital Image Processing Lecture 8: Image Enhancement in Frequency Domain II Naveed Ejaz.
Spatial Filtering (Chapter 3) CS474/674 - Prof. Bebis.
Image Enhancement in the Spatial Domain.
Medical Image Analysis
Miguel Tavares Coimbra
Image Subtraction Mask mode radiography h(x,y) is the mask.
Image Pre-Processing in the Spatial and Frequent Domain
ECE 692 – Advanced Topics in Computer Vision
Spatial Filtering - Enhancement
Digital Image Processing
Image Enhancement in the Spatial Domain
Lecture 3 (2.5.07) Image Enhancement in Spatial Domain
Linear Operations Using Masks
Image Enhancement in the Spatial Domain
Presentation transcript:

2007Theo Schouten1 Enhancements Techniques for editing an image such that it is more suitable for a specific application than the original image. Spatial domain: g(x,y) = T[f(x,y)] Frequency domain: g(x,y) = FT -1 [H(u,v) F(u,v] = h(x,y)  f(x,y)

2007Theo Schouten2 Point processing Gamma transformation: s = c r 

2007Theo Schouten3 Histogram With histogram equalization we search for a T(r) that makes the histogram as smooth as possible. The T(r) that accomplishes that is: s k = round( L  j=0 k (n j / n) ) with n k the number of pixels with gray level k, n the total number of pixels and L the number of gray levels. Landsat image river Taag Histogram

2007Theo Schouten4 Examples Original Contrast stretched Hist. equalization Local histogram Local contrast Local contrast enhancement: g(x,y) =  (x,y) + kM (f(x,y) -  (x,y))/  (x,y)

2007Theo Schouten5 Smoothing This is used for the blurring of an image: the removal of small details and the filling in of small gaps in lines, contours and planes, and also reduces the noise in an image. In the frequency domain smoothing becomes: G(u,v) = H(u,v)F(u,v) : low pass filter In the spatial domain smoothing is the removal of drastic changes by averaging the gray levels in a certain region with a positive weight.

2007Theo Schouten6 Smoothing Frequency domain Butterworth LPF: H n (u,v)=1/(1+(  (u 2 +v 2 )/D 0 ) 2n ) the Exponential LPF: H n (u,v)=exp(-  (u 2 +v 2 )/D 0 ) n ) the Gaussian LPF: H(u,v)=exp( - (u 2 +v 2 ) / 2 D 0 2 ) Ideal LPF, the rings of especially the rivers can clearly be seen. Right image : Butterworth LPF with n=5. Here the ringing has decreased.

2007Theo Schouten7 Gaussian LPF

2007Theo Schouten8 Smoothing spatial domain A linear filter can be shown as a convolution mask: | | | | | | | | | | | | | | | | (1/25)| | (1/21)| | (1/81)| | (1/256)| | | | | | | | | | | | | | | | | | The 2 right filters are examples of separable filters, they can be executed as a convolution with (1/9) | | respectively (1/16) | | in the x direction, followed by a convolution in the y direction. The right filter is a poor approximation of the Gaussian function g(x,y) = c exp( (x 2 +y 2 ) / 2  2 ), better ones are not separable. With a non-linear "rank" or "order-statistics" filter, the pixel values in the neighborhood are sorted according to increasing value, the value at a fixed position in the row is chosen to replace the central pixel. Choosing a value in the middle results in a so-called median filter.

2007Theo Schouten9 Examples mean and median image average filter median filter Original 5x5 mean 5x5 median 20% impulse noise 5x5 mean 5x5 median

2007Theo Schouten10 Noise models

2007Theo Schouten11 Noise models (2)

2007Theo Schouten12 Sharpening This is used to bring fine details to the front of the image and to sharpen the edges of objects. High Pass Filter: G(u,v) = H(u,v) F(u,v). ideal HPF: H(u,v) = 1 if  (u 2 +v 2 ) > D and 0 otherwise, see fig Butterworth HPF: H(u,v) = 1/(1+D/  (u 2 +v 2 ) ) 2n, see fig Exponential HPF: H(u,v) = exp(- D/  (u 2 +v 2 ) ) n Gaussian HPF: H(u,v) = 1- exp( - (u 2 +v 2 ) / 2 D 0 2 ) see fig. 4.26

2007Theo Schouten13 Examples combinations Here an example from High Frequency Emphasis, where the third order Butterworth HPF is added to the original image using the proportion 0.7:1. Blur masking: I-LPF(I) The impact of fragments from the Shoemaker- Levy comet on Jupiter on July 19th 1994.

2007Theo Schouten14 Sharpening in the spatial domain Differences in the gray levels between pixels, often as an approximation of the derivative of the image function f(x,y):  f(x,y) = (  f/  x,  f/  y )  ( (  f/  x) 2 + (  f/  y) 2 ) results in the gradient image. In the simplest case, one discretely approximates:  f/  x = f[x,y] - f[x-1,y] g( f[x,y] ) = |  f/  x | + |  f/  y | Other manners often used to determine the derivative are: |1 0| | 0 1| | | |-1 0 1| | | |-1 0 1| |0 -1| |-1 0| | 0 0 0| |-1 0 1| | 0 0 0| |-2 0 2| | 1 1 1| |-1 0 1| | 1 2 1| |-1 0 1| Roberts Prewitt Sobel Prewitt and Sobel take more pixels into account and are thus less sensitive to noise.

2007Theo Schouten15 Sobel Original Simple derivative Sobel Gausian

2007Theo Schouten16 Laplacian Other sharpening operators are derived from the Laplacian:  2 f(x,y) =  2 f/  x 2 +  2 f/  y 2 Discretely, one can use masks: | | | | (1/9) | | or (1/5) | | | | | |

2007Theo Schouten17 Moon

2007Theo Schouten18 Whole body bone scan