CH 3 Image Enhancement in the Spatial domain

CH 3. Image Enhancement in the Spatial domain 3. 1 Background 3. 2 Grey level transformations 3. 3 Histogram processing histogram equalization, histogram matching local enhancement 3. 4 Enhancement using A/L operations 3. 5 Basic spatial filtering 3. 6 Smoothing spatial filters 3. 7 Sharpening spatial filters 3. 8 Combining spatial enhancement methods 1

3. 4 Enhancement using A/L operations Masking- selecting subimages in an image - Referring to as ROI (region of interest) processing - In term of enhancement, isolating an area for processing Arithmetic operations: Addition, Subtraction, Multiplication: used to implement gray-level rather than binary Division: Logic operations: And: used for masking (Fig. 3. 27) Or: used for masking Not operation: negative transformation Also are used in conjunction with morphological operations 2

Enhancement using arithmetic/logic operations 3

3. 4. 1 Image subtraction The difference between two images f(x, y) and h(x, y): g(x, y) = f(x, y) - h(x, y) Enhance the difference part of two images Contrast stretching transformation—useful for evaluating the effect of setting to zero the lower-order planes (Fig. 3. 28(d)) 4

Enhancement using arithmetic/logic operations 5

Enhancement using arithmetic/logic operations E. g. Image subtraction: Mask mode radiography (Fig 3. 29) g(x, y) = f(x, y) - h(x, y) : h(x, y) is the mask 6

3. 4. 2 Image averaging 7

Enhancement using arithmetic/logic operations 8

3. 5 Basics of Spatial Filtering 9

Basics of Spatial Filtering 10

Chapter 3 Intensity Transformations and Spatial Filtering

Chapter 3 Intensity Transformations and Spatial Filtering

3. 6 Smoothing Spatial Filters q Used for blurring and for noise reduction. § Blurring is used for removal of detail and bridging of small gaps in lines or curves. Smoothing linear filters q Averaging filter (low pass filter) § Replace the value of every pixel by the average of the gray levels in the neighborhood by the filter mask Reduce sharp transition (such as random noise), : side effect - blurring edges § E. g. The average of the gray levels in the 3 x 3 neighborhoods 13

3. 6. 1 Smoothing linear filters § (a) Box filter: coefficients = 1 § (b) Weighted averaging: give more weight to some pixels 14

Smoothing Spatial Filters Increasing filter mask size =3, 5, 9, 15, 35 15

Smoothing Spatial Filters 16

3. 6. 2 Order Statistics filters (rank filters) u Nonlinear spatial filter based on ordering (ranking) • Median filter § Remove impulse noises (salt and pepper noises) § Represent 50 th percent of a ranked set § Large clusters are affected considerably less • Max filter (100 th percentile filter) useful in finding the brightest points • Min filter (0 th percentile filter) 17

Smoothing Spatial Filters 18

3. 7 Sharpening Spatial Filters Sharpening spatial filter • Highlighting fine detail or • Enhancing detail that has been blurred § Application ranging from electronic printing and medical imaging to industrial inspection § Can be accomplished by digital differentiation 19

3. 7. 1 Foundation § Sharpening filter based on first- and second-order derivatives § Definition for first derivatives in terms of differences • Must be zero in flat segment • Muse be nonzero at the onset of a gray level step or ramp • Must be nonzero along ramps § 20

21

Sharpening Spatial Filters § Definition for second derivatives: is better suited than the firstderivative for image enhancement • Must be zero in flat areas • Muse be nonzero at the onset and end of a gray level step or ramp • Must be zero along ramps of constant slope 22

Sharpening Spatial Filters 23

§ Properties of first derivate: • produce “thick” edges • has a strong response to gray-level step § Properties of a second order derivate: • a stronger response to fine detail such as thin line and isolated points • produces a double response to a gray-level step change 24

3. 7. 2 Use of 2 nd Derivatives for Enhancement - Laplacian Development of the method (Laplacian) A function of f(x, y) of two variables is defined as Filter mask used to implement the Laplacian (Fig. 3. 39) ◦ Various versions: Diagonal, the sign of center value

Image Enhancement in the Spatial Domain

To overcome the shortcoming of the operation ◦ Enhance small detail and preserve background tonality “recover” background features while preserving the sharpening effect By adding the original f(x, y) and Laplacian images (Fig. 3. 40) A negative center coefficient--subtract sharpen result A positive center coefficient—add sharpen result

Sharpening Spatial Filters 28

29

Simplification (Laplacian filtering step + adding original image step one step) • Composite Laplacian mask - no diagonal neighbors • Diagonal neighbors—sharper than no diagonal neighbors 30

Sharpening Spatial Filters 31

Unsharp masking and high-boost filtering Un-sharp masking: The dark room photography ◦ subtract a blurred version of an image from the image itself high-boost filtering: A further generalization of un-sharp masking

Un-sharp Masking and High-boost Filtering The center coefficient of the Laplacian mask: negative or positive High-boost filtering is used when the original image is blurred and dark. 33

34

35

Un-sharp Masking and High-boost Filtering 36

3. 7. 2 Use of 1 st Derivatives for Enhancement - Gradient First derivatives in image processing are implemented using the magnitude of the gradient. 37

Use of First Derivatives for Enhancement Gradient Robert operators Sobel operators Roberts operator Gx = (z 9 -z 5) and Gy = (z 8 - z 6) Sobel operator Gx = (z 3+2 z 6 +z 9) - (z 1+2 z 4+z 7) and Gy = (z 7+2 z 8+z 9) - (z 1+2 z 2+z 3) 38

Use of First Derivatives for Enhancement Gradient 39

3. 8 Combining Spatial Enhancement Methods 40

Combining Spatial Enhancement Methods 41

Summary 3. 4 Enhancement using A/L operations ◦ Image subtraction, Image averaging 3. 5 Basic spatial filtering ◦ Convolution, correlation, mask, kernel etc. 3. 6 Smoothing spatial filters ◦ Linear filter(averaging filter), Order-static filters (median, max, min filters) 3. 7 Sharpening spatial filters ◦ 2 nd derivative (Laplacian), Unsharp masking, high -boost filter, 1 st der. (Robert, Sobel) 3. 8 Combining spatial enhancement methods