Image Processing
- Benedict Häfeli
Color Adjustment
before | after |
|---|---|
|  |
Color adjustment can make the best of your images by brightening up shadows, reducing glare and evenly distribute the images contrast over the available color space, leading to more vibrant coloration.
Color adjustment allows you to configure three parameters:
Contrast:
A sigmoidal contrast factor that gets applied to smooth out the curve in dark and bright areas of the image. It’s a rather abstract value, so you’ll have to find out what works best for your image. values from 20 to 30 are considered very high, though.
Black Point and White Point:
The black and white points are percentage values of the images color range. Setting a black point of 0.02 means that the darkest 2% of pixels in the image will be turned black. Setting a white point of 0.97 means that the brightest 3% of pixels in the image will be turned white. After that, the algorithm will scale the remaining colors accordingly to distribute them evenly over the new color space.
Identity Masking
Identity masking is used to render certain areas of an image unrecognisable. The yellow cloud can apply pixelization or blur directly to images you intend to publish. To configure it can adjust the strength parameter, but you also have to upload upload a mask that defines the areas to be pixelized.
The strength determines how strongly the image is affected. For pixelisation, it is a factor the individual pixels get scaled up by, for blur, it determines the radius of the “sponge” that is wiped over the image. Note that the exact effect will vary depending on the resolution of your input image. Larger images may require higher strength than smaller images. Guideline defaults are a strength of 10 for pixelisation, and 5 or 6 for blurring. This should render the masked parts of the image sufficiently unrecognisable in most cases.
The other thing identity masking requires is, obviously, a mask. Which you can upload from your computer with the provided controls, but you will have to create it yourself. A mask is an image that defines in which areas the operation should be applied. Expected is a jpg image with a black background, in which the areas you wish to be pixelized are marked in white. Here is a short tutorial on how to make such a mask using GIMP, a free image editing program. You may of course use whatever image editing program you prefer to create the image.
How to make an identity mask (with GIMP)
First off it’s a good idea to download an image from the table you want to pixelize. While the pixelization mask does not need to match the resolution of the source image (and for very large images really shouldn’t), it is the most convenient way to create the mask on top of the image you want to pixelize. If the mask has a different aspect ratio than the image it will still work, but will be scaled non-uniformly to match the images ratio. This can make it hard to predict which areas of the image will be masked exactly, so it’s not recommended.
Once you’ve opened the image in GIMP, here’s what you do:
Create a new layer:
In the layers dialogue, click on the left-most icon in the bottom. Enter some name for the layer, and click OK.
In case you don’t see the layers dialogue, you can bring it up by pressing ctrl-L, or by going through the top menu: Windows → Dockable Dialogues → Layers.
Mark the area you want to render unidentifiable:
Activate one of the three selection tools in the toolbox. There’s rectangles, elipses, and free polygons. I’m going to use the polygon for this tutorial. If you don’t see the toolbox anywhere, you can bring it up by pressing ctrl-B or from the top menu, going through Windows → New Toolbox.
Now mark out the area you want to pixelize with the selection tool of your choice. Make sure the layer you created is selected (it should be if you didn’t do anything else in the meantime)! In case of the polygon tool, we can just create arbitrary points to encircle the desired area. To close the polygon just connect back to the first point.
You can select multiple areas at once using the control key if you need to.
Color the mask:
In the toolbox, select the bucket fill tool. Also make sure that your foreground color is set to white. If it isn’t, change it.
With the bucket tool selected, just click inside the area you selected. The result should look something like this:
Almost done, now we merely need to turn everything else black. Click on the lowest layer to select it, and make sure your background color is set to black:
Then just press ctrl-A to select the entire background layer, and hit the delete key. voila:
Export your mask and upload it to the yellow portal:
In the top menu, go to file → Export as:
In the menu that pops up, enter a filename (with a .jpg extension), and click export. Another menu with JPG options will pop up. You can put the quality a decent way down, we don’t need it here. Also uncheck all checkboxes except the lowest one, then click Export:
All that’s left now to make your identity mask work is to go back to the yellow portal, click the “choose” button in the upload widget, select the file you just created, then click on upload. And finally, of course, save your table configuration. The next image coming in should already have the pixelisation applied.
Should you ever need to change the mask, you can upload another one at any time. This will simply overwrite the mask that was used for this table. Masks uploaded for one table do not affect other tables.