No announcement yet.

Feature suggestion: Image diff segments

  • Filter
  • Time
  • Show
Clear All
new posts

  • Feature suggestion: Image diff segments

    Hello everyone, long time user/lurker, first time poster

    I've been using Beyond Compare since around the 2.x days. I realize the Image diff component has been overhauled in 4.x but to be honest, it's just as useless (for me, at least) as the 3.x version.

    I would like to suggest a way that two images can be segmented and the image segments compared.

    Here's a real world use-case: I take a screenshot of two web forms with a bunch of input fields. They're almost identical, but one has 20 fields the other has 21 fields - the only difference is an additional field injected after the 10th field. Now all of a sudden the diff is completely broken because it will show the entire image as different below the 10th field.

    My proposal is to be able to define (or ideally automatically) have horizontal diff segments defined for the left side and again for the right side.

    To better illustrate my thoughts, I've attached a screenshot.

    This is what I get currently - completely useless to me (sorry, a bit harsh). There should be a color overlay on each image, too (like in my mockup above):

    The fields do line up after the first few, but because I can't segment the image, it doesn't work for me. I think this is a pretty common use-case, and this is a feature I have been hoping for for many years.

    So ultimately here's what would need to happen to make it work:
    • Color overlay (ex: blue overlay on left, red overlay on right, slighly opaque so when something matches its red on blue = purple)
    • Ability to define (or it would be sweet to auto-find) diff segments

    Thanks for hearing me out!
    PS: is there a BeyondCompare 4.x plugin API planned? I would be interested in implementing this myself if possible...

  • #2

    Thanks for the feedback and suggestions. Given this screenshot, might OCR software work better? This picture appears to be mostly text data, which could be interpreted and then compared in the Text Compare; this could be done with an external conversion format.

    BC4 does not have the BC2 plugin structure currently enabled. We are adding a new featured called External Formats, which is a new format type which can call any external application. Instead of a plug-in, this format type could reference any other program.
    Aaron P Scooter Software


    • #3
      Thanks Aaron,

      I think in this case doing an OCR generated diff would be quite helpful - but if the UI was anymore complex, it would get tricky. I could dump the text to a text file and then do a text diff with Beyond Compare. In this particular case I'm trying to visualize the difference between two of our UIs and send it to a customer. The screenshot from Beyond Compare wasn't practical to send, but I did the one in Photoshop which took way longer than I wanted lol.

      I think with the combination of the existing tolerance setting and adding the ability to split an image and compare left segments to right segments would be very useful. I suppose I could split them up in photoshop and then compare each segment in Beyond Compare, but that could be a lot of work and it just gets more cumbersome with the more segments.

      Even if it was just 1-dimenionsal segments, I think it would be very useful still. For example, the user can only segment out horizontal chunks or vertical chunks.

      Regardless of the segments, I also think the slightly transparent color overlay is very helpful, too - at least with lighter-colored images, and that might be useful as just a general feature to the image diff stuff

      Anyway, I'm just kinda repeating what I said before . I've been a huge fan of you guys forever, and this is really the one component that I seem some room for improvement on. You guys rock!

      Thanks for hearing me out!


      • #4
        Thanks. I'm not sure if we'll be able to meet this type of requirement, but I'll add your notes to our Customer Wishlist. Given the example, I'd still recommend an OCR software to help with this specific case.
        Aaron P Scooter Software


        • #5
          Hi Aron,

          Have you got the OCR software capabilty incorporated to BC4 now, if so which version.


          • #6
            No, we do not incorporate OCR software, but if you have a command line utility which can perform the conversion it can be plugged into a Text format conversion.

            We have a KB article detailing how to incorporate a custom program, here:
            Aaron P Scooter Software