Image background removal algorithms - algorithm

I've lots of images with a white background and I need to remove it (make it transparent). I've already tried imagemagick convert and for some image that worked quite well, but still cannot be used on dark background.
I'm kind a noob in this branch of programming... Someone know some good algorithms to remove the background?
Plus, would be great if it could correctly process an image like this:
http://dev.addvert.it/cache/24eeab00e5987452d09fbeec0c7678d6_w472_h472_sc.jpg
The problem with convert diff is it erase a lot of the image, while the border algo doesn't touch the central part.
I totally understand how hard is for a calculator to do something like this and I'm pretty amazed with the results of convert, but if there's the chance of a better solution, why not ask? :D

You are not constrained to converting white pixels to transparent. You could maybe take the colour of the top-left corner pixel and make all pixels that colour transparent and it will work for black backgrounds too:
As suggested by #emcconville, the first of the following two options is more succinct:
convert towel.jpg -fill none -fuzz 2% -draw 'matte 0,0 replace' output.png
or, per my original,
convert towel.jpg -alpha on -fill none -fuzz 5% -draw 'color 0,0 replace' output.png
You may find this more readable...
convert towel.jpg -fuzz 5% \
-transparent $(convert towel.jpg -format "%[pixel:p{0,0}]" info:) \
output.png
In the second line, $(convert ...) just gets the colour of the top-left pixel and feeds that into the middle of the outer convert command as the colour to make transparent - but it is only doing the same thing as the first version.

Related

I want to remove the white background from and image not inside just outside the colored pixels using convert command in terminal

I've image an image with white background. I used command to make it transparent convert imoji.png -fuzz 20% -transparent white result.png I got the result.png. The command has removed pixels inside the emoji and some other parts. I need it to be something like this. I made it using photoshop by reducing the tolerance of the magic wand tool. Help me to do the same using the convert command in the terminal. I reduced the -fuzz 20 to 1% still not getting the result.
Use flood fill from the top left corner of the image in ImageMagick.
convert imoji.png -fuzz 20% -fill none -draw "matte 0,0 floodfill" result.png

How to convert coloured Captchas to Grey Scale?

I'm trying to make a capcha solver, but I have ran into some trouble. The captcha that I am trying to solve has different coloured backgrounds.
I need to convert it to black text on white background so that it could easily be recognised by tesseract-ocr
I have tried
convert *.png -threshold 50% *.png which only shows some of the digits.
The problem with simple 50% thresholding is that both colours may be lighter than 50% grey and will therefore come out as white. Or, conversely, both colours may be darker than mid-grey and therefore bith come out as black.
You need to do a 2-colour quantisation to get just 2 colours, then go to greyscale and normalize so the lighter colour goes white and the darker one goes black. I am not near a computer, to test, but that should be:
convert input.png -colors 2 -colorspace gray -normalize result.png
Now, you will find some images are inverted (black on white instead of white on black), so you can either test the top left corner pixel and if it is white, then invert the image. Or, you could get the mean of the image and if it is more than 0.5 that would indicate that the image is largely white and therefore needs inverting.
Invert with:
convert input.png -negate output.png
Get top-left pixel with:
convert image.png -format '%[pixel:p{0,0}]' info:-
Get mean value with:
convert image.png -format "%[mean]" info:-

Making half an image transparent from the command line

I don't really know where to start with this one. I'm trying to do something that I thought would be relatively simple to accomplish with imagemagick, but I don't know the exact command to start with. I need to draw a line through an image, and then make everything above the line transparent in the image, and make everything below the line, the orignal image. What would be the best way to accomplish this using imagemagick?
So what I've come up with for now is to crop the image, and then resize it to the original size, but with a transparent background. The command I use is this, but it always comes out black. I'm not understanding why.
convert -background none -gravity south out.png -resize 400x200 -extent 400x400 result.png
Thanks for all of the help!
Here's a fairly easy way to do it. First, enable an alpha channel in case your image doesn't have one, then select the alpha channel for modification by the following -fx command. In there, if the current j is greater than half the height of the image, make the alpha layer opaque, else transparent. Easier than it sounds!
So, using this start image:
convert bean.jpg -alpha on -channel A -fx "j>h/2?1:0" result.gif
Or, the other way:
convert bean.jpg -alpha on -channel A -fx "j<h/2?1:0" result.gif
Or the other, other way:
convert bean.jpg -alpha on -channel A -fx "i<w/2?1:0" result.gif
Or, if you are feeling particularly silly on a Friday morning...
convert bean.jpg -alpha on -channel A -fx "hypot(i,j)/400-0.8" result.gif

Mogrify / ImageMagick change jpeg background from grey to white

we run a mogrify batch job like mogrify -path "WEBREADY\DONE" -layers merge -trim +repage -resize "1200x1200>" -define jpeg:extent=500kb -format jpg *.*
This works fine for images with a white background. But recently we saw a 2% error rate and discovered the error. Mogrify trims as expected only on the web the images is surrounded by white so this looks quite silly with grey-ish in corners.
THe solution would be to either
Detect the non-white background. And in that case -only- do not trim
Or to detect the background color (it is JPG) in top left corner and forst set this to an all white background.
Can this be done in my 1 command above for the commandline? (so find the grey background first and change it to white)
You can get the colour of the top-left pixel like this:
convert shirt.jpg -crop 1x1+0+0 txt:
# ImageMagick pixel enumeration: 1,1,65535,srgb
0,0: (57311,57311,57311) #DFDFDF srgb(223,223,223)
So, it is #DFDFDF.
If you now try and replace that shade with yellow (I know you want white but you can't see that on here):
convert shirt.jpg -fill yellow -opaque "#DFDFDF" result.png
and you can now see the problem. The background is not uniform and its colour also appears in your shirt. If you add some fuzz, it makes it more pronounced:
convert shirt.jpg -fill yellow -fuzz 10% -opaque "#DFDFDF" result.png
Even if you try flood filling from the top-left, you still have to hope that there are no pixels in the edges of your shirt that are similar to that corner:
convert shirt.jpg -fill yellow -fuzz 5% -floodfill +0+0 "#dfdfdf" result.png

To remove background greyed pixels from image

I want to remove background unnecessary greyed pixels in above image.
May i know how to do that.
Quick and dirty with ImageMagick:
convert 9AWLa.png -blur x1 -threshold 50% out.png
Somewhat better, still with ImageMagick:
convert 9AWLa.png -morphology thicken '1x3>:1,0,1' out.png
Updated Answer
It is rather harder to keep the noisy border intact while removing noise elsewhere in the image. Is it acceptable to trim 3 pixels off all the way around the edge, then add a 3 pixel wide black border back on?
convert 9AWLa.png -morphology thicken '1x3>:1,0,1' \
-shave 3x3 -bordercolor black -border 3 out.png
Some methods that come to my mind are
If the backgroud is gray color rather than sparse black dots then you can convert the image in binary by thresholding it with proper value of grayscale. i.e. all values above particular values of pixel are white and all values below that are black. Something like this.
Another thing you can do is first smoothing the picture my some filter like mean or median filter and then converting into binary presented in previous point.
I think in this way the unnecessary background can be removed

Resources