mgedmin
Wishlist: optionally highlight differences with a color
Actually not a big deal. but it difficult to change diff color which is always black. is it possible to add a command line param for that? it might be usefull for black and white images
I assume you're talking about -H? The diff color isn't actually black -- the similar areas have the opacity reduced to 25%, the differing areas remain in the original color.
This is somewhat configurable using the --opacity and --bgcolor options.
Tinting the differing areas sounds like it could be a good idea in some situations. Do you have any sample images you could share that don't look good with what imgdiff currently does?
hi, Thanks for a very quick reply. I used People and People2 images from https://github.com/Huddle/Resemble.js/tree/master/demoassets
as you can see in their diff images the diff is either yellow or pink. That's why I think is a good idea to add an option to change the diff color. not the bg of the image.
Can you take a look?
Thanks.
Thank you for the sample images.
I agree that some of the differences are hard to see with imgdiff, but I'm not very fond of the magenta/yellow in https://github.com/Huddle/Resemble.js/blob/master/demoassets/readmeimage.jpg.
For the record, here's what imgdiff produces:
So it is not possible to change the diff color? I'd say it can improve user experience. a simple paramater can allow them to manipulate diff color.
On Fri, May 22, 2015 at 11:51 AM, Marius Gedminas [email protected] wrote:
Thank you for the sample images.
I agree that some of the differences are hard to see with imgdiff, but I'm not very fond of the magenta/yellow in https://github.com/Huddle/Resemble.js/blob/master/demoassets/readmeimage.jpg .
For the record, here's what imgdiff produces: [image: screenshot] https://cloud.githubusercontent.com/assets/159967/7767141/c3547a04-0078-11e5-8011-ad3dc6024c15.png
— Reply to this email directly or view it on GitHub https://github.com/mgedmin/imgdiff/issues/1#issuecomment-104575983.
Currently it's not possible, in imgdiff.
I'd like it to be, but this needs:
- a design (how should things look?)
- an implementation
I have neither, and I'm unlikely to come up with a solution any time soon.
BTW some interactivity could help. E.g. a slider that lets you adjust the opacity. Or a slider that lets you adjust the hue of the highlighting (once that's been implemented). Or a selectable mode (side-by-side, overlay, half-and-half).
That would be a separate issue, and I also am unlikely to find the time to work on it.
FWIW I like what ImageMagick's compare does (reduced opacity + differences in red). It works great for mostly-greyscale images (e.g. I used it to compare PDFs full of text), but it probably won't work well with colored images like my sample set.
I need a python library whcih can be imported in my testing library. Do you have some links to documentation which inpired you to write this tool?
On Fri, May 22, 2015 at 12:53 PM, Marius Gedminas [email protected] wrote:
FWIW I like what ImageMagick's compare does (reduced opacity + differences in red). It works great for mostly-greyscale images (e.g. I used it to compare PDFs full of text), but it probably won't work well with colored images like my sample set.
— Reply to this email directly or view it on GitHub https://github.com/mgedmin/imgdiff/issues/1#issuecomment-104602147.
I need a python library whcih can be imported in my testing library. Do you have some links to documentation which inpired you to write this tool?
If you're looking for tools to compare images, the words to google are "perceptual diff". I did some research not too long ago: https://github.com/mgedmin/mgp2pdf/blob/master/COMPARE.rst
great. In case nobody told you earlier. THANK YOU for your github activity :)
On Fri, May 22, 2015 at 2:21 PM, Marius Gedminas [email protected] wrote:
I need a python library whcih can be imported in my testing library. Do you have some links to documentation which inpired you to write this tool?
If you're looking for tools to compare images, the words to google are "perceptual diff". I did some research not too long ago: https://github.com/mgedmin/mgp2pdf/blob/master/COMPARE.rst
— Reply to this email directly or view it on GitHub https://github.com/mgedmin/imgdiff/issues/1#issuecomment-104626735.