2 Replies Latest reply on May 6, 2015 2:43 PM by jeromee77

    LR CC. Face recognition with pictures scanned as negatives

    jeromee77

      Hi!

       

      I have scanned a lot of old "negativ" pictures using my D90, connected to my PC.

      I got NEF files (no JPEG). The resulting quality is ok for me at this step.

      As they are B and W negatives, I used a basic development preset (curves) to get the positiv image, and work on them on LR.

      No major issue at this point.

       

      I tried to run face recognition on thoose files. Not very lucky, as I had to go through a lot of pictures to draw the box and enter the person keyword.

      But at the end it worked as it does with all my other folders (that means with good things and sometimes quite strange results, but this is another subject)

       

      To be complete, I'll add that all the thumbnails in the filmstrip at the bottom of the screen are ok (though I had to wait for LR to rebuild the previews, that seems ok for me).

       

      Then, my question:

       

      Looking at the "confirmed" people window, I have only thumbnails of the negativ pictures.(see screen capture beow)

      I think they are coming from the original Nef file (negativ pictures), and do not take care of the inversion I did in the development module.

       

      Checking faces on negatives views is quite unusal for most of people... And I can't do that no more (I have worked for years with negativs pictures in a dark room, but that is years ago...)

       

      Of course, I could try to export all thoose pictures, and the re import them (I'll get positiv pictures at this step).

      Unfortunately, I have thousands of that kind of pictures, tagged with keywords, GPS locations and so on....

      And this will not allow me to do fine work on them (I'd like to keep the RAW files...).

       

      So: is there a way to get (in the "confirmed" people window) thumbnails looking as "positives", that means after the treatment I did in the development module?

       

      I'll apreciate help and advices.

       

      Thanks

       

      Jérôme

      France.

       

      Runing LR CC 2015 on W7, i7 processor, 8 gig RAM.

       

      LR CC 060515.jpg

        • 1. Re: LR CC. Face recognition with pictures scanned as negatives
          trshaner Adobe Community Professional & MVP

          I'm don't think there is anyway to invert the thumbnails used by Face Recognition, but there is a better option.

           

          Why not use Export or 'Edit in PS' to convert the camera raw images to 16 bit ProPhoto RGB TIFF. Next invert the image files in PS and convert them to Grayscale mode. You can use Batch processing to apply these steps for a group of image files. Then save the image files and process them inside LR. The file size is not outrageous since it is Grayscale and you can apply ZIP TIFF compression to reduce it further. You can also apply special processing unavailable in LR  using PS (Content Aware Fill,etc.) to the "base" TIFF file.

           

           

          This article may be of interest concerning processing color negative images. Many of the techniques used apply to B&W negative processing as well.

           

          https://luminous-landscape.com/scannerless-digital-capture-and-processing-of-negative-film -photographs/

           

          A PDF with full workflow-details is available in the above article.

          • 2. Re: LR CC. Face recognition with pictures scanned as negatives
            jeromee77 Level 1

            Thanks for your advice and for the link to Luminous-Lanscape. I know this place and often go there to get informations.

             

            I downloaded the PDF file: it seems very interresting. I'll try to read that in details tomorrow (French time... )..

             

            And I'll try the process with few pictures to get an idea of the results, and try to know how long It'll take for my thousands of pictures I have to work on!

             

            Thanks again.