Legacy ImageMagick Discussions Archive

How do i re-position images when i use +append?
How do i move those images over that 500x500 canvas?
I can use -composite. How do i do it with append?

See -swap at http://www.imagemagick.org/Usage/basics/#swap. You can also just list your images in the order you want them appended. There is no need for -size 500x500. The output of append will be as large as needed to hold your images.

See http://www.imagemagick.org/Usage/layers/#append. Append just puts the images together. It has no size control. Any mismatched sizes will be filled with background color. If you want to put images in special places in a 500x500 background image, then you need either -compose over -composite (pair-wise) or -page ... -flatten. See http://www.imagemagick.org/Usage/layers/#flatten

I think this is a follow on post from this thread fmw42

I would probably use -layers and there are some examples on this page

Yes sir. But question will be useful to others. I asked here.
Thanks for layers link. It is working good.

Is there any face detecting algorithm in imagemagick?

Not that I know of; but there was a thread about it a couple of years ago. I tried it out with php and an external library but I have formatted my PC since then and no longer have the code.

I have had a look through the forum and I based my code on: http://liuliu.me/ccv/js/nss/

Bonzo wrote:I have had a look through the forum and I based my code on: http://liuliu.me/ccv/js/nss/

What was the outline of the methods used in your IM code or do you have a link to your IM code.

I have deleted the code from my server and PC.

As far as I can remember:
1/ Send the image from the webpage to php via Ajax and base64
2/ Save a temporary file for the next stage. If you can remember back a few years I was having a problem with base64 and so I had to do this step with GD as I could not get IM to work. If I could have got the base64 to work I could probably have avoided the temporary image.
3/ Cropped out all the heads with IM and saved each one.
4/ Draw a red square outline around each head with IM, resize and save.

It worked quite well on my sample images but I did not take it any further. The main idea was to generate a webpage with the main image and heads outlined then a list of heads with a space next to it. I was then going to supply links to relatives and they could then use an online form to fill in the missing names or supply alternative names if they thought any were incorrect.

The question is how did you detect where the head areas were in the image.

I used http://liuliu.me/ccv/js/nss/ and it gave me the co ordinates