Image processing

IMAGE PROCESSING

Presented By :-

Lovely Soni

Chartered Institute of Technology, aburoad.

WHAT IS IMAGE ?Image is an Artifact that reproduce the

likeness of some subjects usually a physical object.

IMAGE ENHANCEMENT Improving Quality of Image Two types of Work Area:

I) Special DomainII)Frequency Domain

In Special Domain enhancement achieved by processing on the actual pixels of image.

In Frequency Domain it is achieved by processing on a frequency or intensity.

HISTOGRAM

HISTOGRAM is graph of color intensity value v/s no. of pixels with that intensity in image.

HISTOGRAM is Blueprint of image. How to construct HISTOGRAM? Working with RGB image of 24 bit image ,

HISTOGRAM is prepared for each color . Working with Grayscale image of 8 bit

maximum 256 shades ,single HISTOGRAM is prepared.

FILTERS FILTERS FOR IMAGE PROCESSING: Brightness Filter Smoothing Filter Noise Reduction Filter Contrast Filter

Brightness Filter To increase BRIGHTNESS of image, move its HISTOGRAM

towards high intensity. It is achieved by adding some fixed value in each pixels

of image. It means HISTOGRAM is shifted by that value and

BRIGHTNESS increased by the same.

BRIGHTNESS FILTER For eg. To increase BRIGHTNESS in RGB image, we have

to add K value in each color that is, RED, GREEN & BLUE. Processed pixel= source pixel + Brightness constant

Brightness Filter

Smoothing Filter To apply SMOOTHNESS, mask of 9 pixels is used

Target pixel

This mask is applied on all pixels to smooth image. To apply this mask we have to move from upper left corner to lower

right corner one by one each pixel. The Target pixel is replaced by the MEAN of mask pixels.

Smoothing Filter

Noise Reduction Filter In image any spike (point or itch) comes during scanning called NOISE. To remove NOISE, in some methods, mask of 5 pixels of “ + “ sign

shape is used. Target pixel

This mask is applied on all pixels of image. Image results in reduced image or low noise image. To apply this mask we have to move this mask from upper left corner of

image to lower right corner one by one on each pixel. The MEDIAN of mask pixels replaces the Target pixel.

Noise Reduction Filter

Contrast Filter To increase CONTRAST of image, spread up its HISTOGRAM over X-axis. It means HISTOGRAM is spread up by a constant value and

CONTRAST increased by the same. For eg. To increase CONTRAST in Gray image, for each pixels , we

have to subtract mid intensity from pixel intensity then increase or decrease it by contrast constant and then add mid intensity value to it.

Processed pixel = (source pixel-128) *(( contrastpercent+100)/100) +128

Contrast Filter

Operations• Euclidean geometry transformation• Color correctness • Digital compositing• Image registration• Image recognition• etc…

Applications• Computer Vision• Face Detection• Feature Detection• Medical Image Processing • Remote Sensing• etc…

Thank You….