15
FILTERING AND MASKING -AMUDHINI.R 111EC102

Filtering and masking

Embed Size (px)

Citation preview

Page 1: Filtering and masking

FILTERING AND MASKING

-AMUDHINI.R111EC102

Page 2: Filtering and masking

MASK

• A mask is a small matrix whose values are called weight.

• Each mask has an origin ,which is usually one of its positions

• Symmetric mask• Non symmetric mask

Page 3: Filtering and masking

MASK

• Input image equal to output image.• Types of maskConvolutionCross correlation

Page 4: Filtering and masking

CONVOLUTION

Mask is placed on the top of the imageMask input image pixel value multiplied with

mask weighs.summed together to yield a single output

value that is placed in the output image at the location of the pixel being processed on the input

Page 5: Filtering and masking

CONVOLUTION

Page 6: Filtering and masking

CROSS CORRELATION

• Without flipping mask is converted to image.• measure the similarity between images or

parts of images.• Mask symmetric – correlation and convolution

same.

Page 7: Filtering and masking

Cross correlation

Page 8: Filtering and masking

FILTERING

• LINEAR FILTERo have the property that the output is a linear

combination of the inputs• NON LINEAR FILTERo Erosion & dilation

Page 9: Filtering and masking

Smoothing filter

• Low pass filter• Noise reduction & image blurring• Removes the finer details of image• Types of filterMean filterGaussian filterMedian filter

Page 10: Filtering and masking

Mean filter

• Averaging filter.• Positive element in

mask.• Size of the mask

determines the degree of smoothing.

Page 11: Filtering and masking

Gaussian filter

Page 12: Filtering and masking

Median filter

• Used to remove the salt and pepper noise

Page 13: Filtering and masking

Sharpening filter

• emphasize the fine details of an image .• Points of high contrast can be detected by

computing intensity differences in local image regions.

Page 14: Filtering and masking

Sharpening using derivatives

• Computing the derivative of an image has as a result the sharpening of the image.

• The most common way to differentiate an image is by using the gradient.

• Using gradient with finite difference has efficient mask.

Page 15: Filtering and masking