Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel...

39
Lecture # 19 Image Processing II

Transcript of Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel...

Page 1: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Lecture # 19

Image Processing II

Page 2: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Image Processing IIImage Processing II

Page 3: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

2 Classes of Digital Filters

• Global filters transform each pixel uniformly according to the function regardless of its location in the image

• Local filters transform a pixel depending upon its relation to surrounding ones

Page 4: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Global Filters: REVIEW

• Brightness and Contrast control

• Histogram thresholding

• Histogram stretching or equalization

• Color corrections

• Inversions

Page 5: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Local FiltersLocal Filters

Blurring

• Sharpening

• Unsharp Masking

• Edge and line detection

• Noise filters

Page 6: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Blurring AlgorithmBlurring Algorithm

For (every pixel in column x)

{ For (every pixel in row y of column x){

average image[x,y] with it’s neighbors;

}

}

Page 7: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Blurring

• Average each pixel with its neighbors by :

multiplying each pixel and its neighbors by 1:

=

- add multiplied pixels and divide by 9

- store result in a new array• The new array contains a blurred image.

1 1 1

1 1 1

1 1 1

Page 8: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Gaussian Blur

Multiply neighbors by less than middle

Page 9: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Blur Demo

•Photoshop

–Filter->Blur->Blur

–Filter->Blur->Gaussian Blur

Page 10: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Local FiltersLocal Filters

• Blurring SharpeningUnsharp Masking

• Edge and line detection

• Noise filters

Page 11: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

• Subtract neighboring pixels by :

multiplying each pixel and its neighbors by:

=

- add multiplied pixels

- store result in a new array• The new array contains a sharpened image.

Sharpening

0 -1 0

0 0

-1-1

-1

+5

Page 12: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Unsharp Masking

• Unsharp masking also sharpens an image• The Algorithm:

1. call the original (unblurred) image A

2. blur the image (call it B)

3. subtract B from the unblurred image (let C = A-B)

4. Multiply C by some number, k (k > 1)

5. Sharpened Image = A + kC

Page 13: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Sharpen/Unsharp mask

• Photoshop– Filter->Sharpen->Sharpen– Filter->Sharpen->Unsharp mask

Page 14: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Local FiltersLocal Filters

• Blurring

• Sharpening

• Unsharp MaskingEdge and line detection

• Noise filters

Page 15: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Edge and Line Detection

• Multiply each pixel and its neighbors by

the following corresponding patterns (numbers)

- add results

- clip to 255

- result =

“Edge” image

-1 0

+1-1

-1 0

0

+1

+1

-1 -1 -1

00

+1 +1

0

+1 Gets Gets

Vertical Edges Horizontal Edges

Page 16: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Edge Detection• edge detection is often used to recognized objects of

interest in the image

Page 17: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Edge Detection

• Photoshop– Filter->Stylize->Find Edges

Page 18: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Local FiltersLocal Filters

• Blurring

• Sharpening

• Unsharp Masking

• Edge and line detectionNoise filters

Page 19: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Noise Filters• A median filter takes away “salt & pepper” noise

(speckles)

Page 20: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Noise Filter

• Photoshop– Filter->Noise->Median Filter

Page 21: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Kernel Filtering

1 1 1

1 1 1

1 1 1

Blur

0 -1 0

0 0

-1-1

-1

+5

Sharpen-1 0

+1-1

-1 0

0

+1

+1

-1 -1 -1

00

+1 +1

0

+1

Edges

Apply same algorithm at every pixel using different kernel values

Page 22: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

ImageImageEditingEditing

Page 23: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Image Editing

Selection Tools

• Painting Tools

• Cut & Paste

• Cloning

• Layers and Blending

Page 24: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Selection Tools

• Lasso Tool- manually draw outline (free hand)

- manually outline (line segments)

- magnetic lasso (semi-automated)

• Magic Wand

DEMOS

ToolBar

Page 25: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Image Editing

• Selection ToolsPainting Tools

• Cut & Paste

• Cloning

• Layers and Blending

Page 26: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Painting Tools

• Airbrush Tool

• Paint Bucket Tool• Paintbrush Tool

• Pencil Tool

DEMOS

Page 27: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Image Editing

• Selection Tools

• Painting ToolsCut & Paste

• Cloning

• Layers and Blending

Page 28: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Cut & Paste

• Word Processors

- cut & paste strings of characters (1D arrays)

• Image Editing

- cut & paste pixels (2D arrays)

- replace old pixels with new pixels

Page 29: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Image Editing

• Selection Tools

• Painting Tools

• Cut & PasteCloning

• Layers and Blending

Page 30: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Cloning• Copy pixels from

one part of an image

- to another part of an image ... Interactively DEMO

Page 31: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Image Editing

• Selection Tools

• Painting Tools

• Cut & Paste

• CloningLayers and Blending

Page 32: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Can create arbitrary number of layers for

- animation

- special effects in movies

- morphing

Layers and Blending

Layer 1

Layer 2

Layer n

Page 33: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

• The idea: Blended image

= .3 x + .7 x

is a weighted combination (sum) of

two or more other images.

Blending

Page 34: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Example Blend

.3 x +.7 x

= Bearastronaut

Page 35: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

• The idea: Create another image

where the value of pixels is the

weighting term for a blend operation:

Masking

Page 36: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Summary

• Digital Images and pixels• Digitize a picture by

– Sampling– Quantization

• Color Models– RGB, CMYK, and HSB

• Storage Formats– .gif, .jpg, .png, .bmp

Page 37: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Summary

• Global Filters– Contrast & Brightness Control– Thresholding– Histogram stretching & equalization– Color corrections– Inversion

Page 38: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Summary

• Local filters– Blurring– Sharpening– Unsharp Masking– Edge and line detection– Noise filters

Page 39: Lecture # 19 Image Processing II. 2 Classes of Digital Filters Global filters transform each pixel uniformly according to the function regardless of.

Summary

• Image Editing– Selection Tools– Painting Tools– Cut & Paste– Cloning– Layers and Blending