16
Presentation on Image processing POOJA G N VIII SEMESTER CIT, Tumkur Mar-15 1

Image processing

Embed Size (px)

Citation preview

Page 1: Image processing

Presentation

on

Image processing

POOJA G N VIII SEMESTER CIT, Tumkur

Mar-15 1

Page 2: Image processing

Mar-15 2

Contents:

Basics of image

Basics of image processing

Basics of image segmentation

Watershed transformation

Point detection

Region based segmentation

Edge detection segmentation method

• First Order Derivative Methods

• Second Order Derivative Methods

• Optimal Edge Detectors

Page 3: Image processing

Image:It is an artifact that depicts or record visual perception.

Image Types:I. Black and white imagesII. Grayscale or intensity imagesIII. Indexed images (intensity images with colormaps)IV. RGB or truecolor images

Image formats: gif, jpg, jpeg, tif, png, bmpMar-15 3

Page 4: Image processing

It is any form of signal processing for which the input is an image, such as a photograph or video frame; the output may be either an image or a set of characteristics or parameters related to the image.

Purpose: Visualization, Image sharpening ,restoration & Image retrieval, Measurement of pattern, Image recognition.

Methods used for image processing are analog and digital image processing.

Mar-15 4

Page 5: Image processing

Mar-15 5

It is the process of partitioning a digital image into multiple segments.

MRI Segmentation:It is an important technique to differentiate abnormal and normal tissue in MR image data.

Segmentation methods:

EDGE DETECTION:

Edge - Area of significant change in the image intensity / contrast.

Use of Edge Detection – Extracting information about the image. E.g. location of objects present in the image, their shape, size, image sharpening and enhancement.

Page 6: Image processing

Mar-15 6

Steps in Edge Detection:

Filtering – Filter image to improve performance of the Edge Detector with respect to noise.

Enhancement – Emphasize pixels having significant change in local intensity.

Detection – Identify edges – thresholding. Localization – Locate the edge accurately, estimate edge orientation.

Methods of edge detection:1. First Order Derivative / Gradient Methods

Roberts Operator, Sobel Operator, Prewitt Operator2. Second Order Derivative

Laplacian, Laplacian of Gaussian, Difference of Gaussian3. Optimal Edge Detection

Canny Edge Detection

Page 7: Image processing

Mar-15 7

First Order Derivative:

At the point of greatest slope, the first derivative has maximum value Gradient: For a continuous two dimensional function Gradient is defined as

Gx= Gy=

Gradient Methods –Roberts Operator: Provides an approximation to the gradient

Gx= Gy=

y

f

x

f

Gy

GxyxfG )],([

)1,(),1()1,1(),()],([ jifjifjifjifGyGxjifG

-1 1

-1 1

1 1

-1 -1

1 0

0 -1

0 -1

1 0

Page 8: Image processing

Mar-15 8

First Order Derivative(contd.,):

Sobel Operator: The 3X3 convolution mask smoothers the image by some amount, hence it is less susceptible to noise. But it produces thicker edges. So edge localization is poor.

Gx= Gy=

Prewitt Operator: It is similar to the Sobel operator but usesslightly different masks.

Gx= Gy=

-1 0 1

-2 0 2

-1 0 1

1 2 1

0 0 0

-1 -2 -1

-1 0 1

-1 0 1

-1 0 1

1 1 1

0 0 0

-1 -1 -1

Page 9: Image processing

Mar-15 9

Second Order Derivative Methods

Zero crossing of the second derivative of a function indicates the presence of a maxima

Laplacian- Defined as

Laplacian of Gaussian-Also called Marr-Hildreth Edge Detector

Difference of Gaussian-

2

2

)2

22(

2

1

)2

22(

22),(

22

21

yxyx

eeyxDoG

Page 10: Image processing

Mar-15 10

Optimal Edge Detector:

Implementation of Canny Edge Detector:1.Noise is filtered out – usually a Gaussian filter is used, width is chosen carefully .2.Edge strength is found out by taking the gradient of the image a Roberts mask or a Sobel mask can be used.

3.Find the edge direction

4.Resolve edge direction

GyGxGyGxG 22

Gx

Gy1tan

Page 11: Image processing

Mar-15 11

Implementation of Canny Edge Detector(contd.,)5.Non-maxima suppression – trace along the edge direction and

suppress any pixel value not considered to be an edge. Gives a thin line for edge.

6.Use double / hysteresis thresholding to eliminate streaking.

Page 12: Image processing

Mar-15 12

Watershed transformation• Watershed model is a mathematical morphological approach.• It transforms image into a gradient image.• Watershed algorithm produces over-segmentation because of noise or textured

patterns.• Traditionally watershed algorithm was applied with median filter to

eliminate noise • it may be good for initial segmentation in a multi -scale resolution as it produces

an over-segmentation. Over-segmentation elimination is also a problem associated with this method which needs further research .

• watershed segmentation using the distance transformEuclidean distance transform of a binary image use IPT function bwdist

• watershed segmentation using gradients• marker-controlled watershed segmentation

Watershed algorithm is new segmentation approach with relatively less application in remote sensing image segmentation than other methods.

Page 13: Image processing

Mar-15 13

Point detection:

Point detection can be achieved simply using the mask below:

Points are detected at those pixels in the subsequent filtered image that are above a set threshold.

Limitation: these methods may provide useful information, but they are

generally limited to relatively simple structures. For complex pathology, more information is often required, which is available in multi-spectral MRI data.

-1 -1 -1

-1 8 -1

-1 -1 -1

Page 14: Image processing

Mar-15 14

Region based segmentation:

Region growing:

Groups pixels or sub-region into larger regions.step1:

Start with a set of “seed” points and from these grow regionsby appending to each seed those neighboring pixels that

have properties similar to the seed.step2:

Region splitting and merging

Advantage:With good connectivity

Disadvantage:Initial seed-points: different sets of initial seed-point cause differentsegmented resultTime-consuming problem

Page 15: Image processing

Mar-15 15

Unseeded region growing:

• no explicit seed selection is necessary, the seeds can• be generated by the segmentation procedure automatically.• It is similar to SRG except the choice of seed point

Advantage:

• easy to use can readily incorporate high level knowledge of the image• composition through region threshold

Disadvantage:• less speed

Application• Muscle Injury Determination

Page 16: Image processing

Mar-15 16

Thank you