by January 1, 2007 0 comments

There are two important things you need to do before working with any image.
First, never work on the original file. Always work on a copy. That way, you do
not lose the original image if something goes wrong. Second, even with the copy,
do not work on the original layer. Create a duplicate layer and do all your work
on it. So, let’s get started.

Open the copy in Photoshop. The picture looks slightly pixelated, with thick
dark horizontal bands all over the image. We have to remove the bands and the
lines and make it brighter and smoother. After the enhancement, we need to
compare the difference between original and enhanced image
To create a new layer, select the image and on the layer palette (this will be
labelled ‘Background’ by default), and drag it onto the Create-New-Layer
icon. The first step is to clean the image, using the Heal Brush and remove
obvious marks and blemishes including the banding. Make the brush size small in
order to sample and brush away the blemishes without losing important details.

Direct Hit!
Applies To:
Graphics professionals
USP: Clean up poor quality images and make them better
Primary Link:
Google Keywords: enhance poor photo

The next step is to remove dust and scratches using the Dust & Scratches
filter. Two slider controls adjust the degree to which this filter affects the
image. Radius determines the size of the dust or scratch that will be
recognized. Threshold sets the difference between the element and surrounding
area. For this image I have set radius to 4 pixels and threshold to 0.

Now we should remove the banding and also give the whole image a creamier
complexion. Gaussian Blur is the best filter for this. Blurring lightens the
pixels found next to the hard edges of well defined lines and shaded areas.
Gaussian Blur smoothens and softens images in variable amounts. Gaussian Blur
radius value is from 0.1 to 250 pixels. For this image we use 0.5 pixels. Next
open the Add Noise filter to give human skin a natural look. Lets use 2.0% noise
with ‘Monochromatic’ selected. Select the skin area using the pen tool to draw
smooth freehand curves. Now, go to the Path window in the layers palette. There
we can see the work path. Select the work path and click on arrow on the right
and select Make Selection from there. In the window that opens, you have to give
the feather radius. Use a value higher than 10. Now the skin area is selected.
Go to Levels or Curves and adjust the color.

Open a copy of the image in Photoshop. This image has both banding and pixelation that we will need to remove


Use the Healing Brush tool to clean up the image. Don’t use the Clone tool for this


Open a copy of the image in Photoshop. This image has both banding and pixelation that we will need to remove


Use the Dust and Scratches filter to remove pixelation from our photograph


Gaussian Blur smoothens the image and gives it a natural creamy look


The picture has become too smooth making it look artificial. Add Noise filter gives the skin a natural tone


On the left is our original photograph. On the right is the cleaned up and enhanced image. The difference is obvious


After correcting the skin color, select the clothing and do the color
correction for it in the same way. If there are different elements of clothing
of varying color or tone, you will have to color correct each separately. You
will notice that the color in the background is spoiled by few patches in some
areas. What you need to do is to remove the patches and make the background
clear or replace it with any other color of your choice.

Finally, we need to sharpen the image little bit. We applied a few filters
and these filters removed the sharpness of the image. We need to compensate for
this. Choose >filter >unsharp mask filter. Apply for this filter according
to the image. Don’t apply more sharpness than the image can take.

That’s all. Compare the original and the final images to see the difference

No Comments so far

Jump into a conversation

No Comments Yet!

You can be the one to start a conversation.