Announcement

Collapse
No announcement yet.

Embed Face Detection for capture in New Database

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Embed Face Detection for capture in New Database

    Over some months, we have spent significant time identifying people in ACDSee.
    I have embedded the metadata hoping that the facial recognition data would be embedded in the individual image files.

    Unfortunately, the database has been giving errors and automatically closes the ACDSee application.
    Consequently I created a new database.

    However, when I created a new database and the facial recognition appears to be running from scratch.
    1. Is there a way that I can get ACDSee to reuse the embedded data from the previous database. – Should I run the CATALOG function on the new database
    2. Where is the Face data stored in ACDSee. Is there any way to confirm that the face data has been embedded, please.
    3. Is there a way of copying the face names from ACDSee into the IPTC/IMM Tags or Keywords fields as a batch process, please.
    Thanks

  • #2
    Export old database into a new one ?

    Comment


    • #3
      Looking at my pictures with XNVIEW MP, a jpg picture, no XMP sidecar file, XMP data embedded into jpg.

      Comment


      • #4
        Originally posted by BBCanb View Post
        ....is there a way that I can get ACDSee to reuse the embedded data from the previous database. – Should I run the CATALOG function on the new database

        If you embedded the ACDSee metadata in your images (Tools/Metadata/Embed ACDsee Metadata] and providing the box for ACDSee Face Data was ticked in the Embed ACDSee Metadata dialog, then the face metadata should be in the images, and can be read back into the new database using Tools/Database/Catalog Files

        Click image for larger version  Name:	Embed dialog.jpg Views:	0 Size:	119.3 KB ID:	59255

        Is there any way to confirm that the face data has been embedded
        When face recognition is run, or rerun on an image it sets the embed metadata pending flag (as does assigning any ACDSee metadata)
        Selecting Embed pending from from the special items in the Manage Mode Catalog Pane.will show any images that have the flag set (ie. Metadata not yet embedded).

        If you view the XMP metadata in ExifTool you will be able to see the face metadata fields themselves..

        Is there a way of copying the face names from ACDSee into the IPTC/IMM Tags or Keywords fields as a batch process
        AFAIK not within ACDSee.

        Edit:
        Experimentally using a command line Exiftool instruction I can batch copy the RegionsRegionListName (face name) to IPTC:Keywords for images that have only a single name. There are some complications that need to be resolved where there are multiple face ID's in an images XMP. metadata, and also perhaps in being able to add the keywords where there are already existing keywords. I haven't explored solutions to those. The other issue is that the IPTC metadata would bave to be specifically read back into the ACDSee database.
        Last edited by Greyfox; 01-22-2022, 06:28 AM.

        Comment


        • #5
          Originally posted by Greyfox View Post
          Edit:
          Experimentally using a command line Exiftool instruction I can batch copy the RegionsRegionListName (face name) to IPTC:Keywords for images that have only a single name. There are some complications that need to be resolved where there are multiple face ID's in an images XMP. metadata, and also perhaps in being able to add the keywords where there are already existing keywords. I haven't explored solutions to those. The other issue is that the IPTC metadata would bave to be specifically read back into the ACDSee database.
          You can do more then just copy the face name from AC namespace in to IPTC namespace: Link

          Comment


          • #6
            Originally posted by Emil View Post
            You can do more then just copy the face name from AC namespace in to IPTC namespace: Link
            Thanks for that link Emil. Interesting. I'm not sure at the moment how to best handle duplicated fields (the same "RegionsRegionListName" field name for each face ID) and how to get the names from those fields to add serially to IPTC:Keywords with "comma" delimiters. If it is adding to other keywords already there, then the first entry may have to include a leading "comma"
            Perhaps when I get some spare time I'll have another crack at it.

            Comment


            • #7
              Originally posted by Greyfox View Post


              If you embedded the ACDSee metadata in your images (Tools/Metadata/Embed ACDsee Metadata] and providing the box for ACDSee Face Data was ticked in the Embed ACDSee Metadata dialog, then the face metadata should be in the images, and can be read back into the new database using Tools/Database/Catalog Files

              Click image for larger version Name:	Embed dialog.jpg Views:	0 Size:	119.3 KB ID:	59255


              When face recognition is run, or rerun on an image it sets the embed metadata pending flag (as does assigning any ACDSee metadata)
              Selecting Embed pending from from the special items in the Manage Mode Catalog Pane.will show any images that have the flag set (ie. Metadata not yet embedded).

              If you view the XMP metadata in ExifTool you will be able to see the face metadata fields themselves..


              AFAIK not within ACDSee.

              Edit:
              Experimentally using a command line Exiftool instruction I can batch copy the RegionsRegionListName (face name) to IPTC:Keywords for images that have only a single name. There are some complications that need to be resolved where there are multiple face ID's in an images XMP. metadata, and also perhaps in being able to add the keywords where there are already existing keywords. I haven't explored solutions to those. The other issue is that the IPTC metadata would bave to be specifically read back into the ACDSee database.
              Thank you. Embedding the metadata allowed the database to pick up the facial recognition. Phew...
              Data is all there.
              Thank you so much.

              Comment


              • #8
                Originally posted by brajaq View Post
                Looking at my pictures with XNVIEW MP, a jpg picture, no XMP sidecar file, XMP data embedded into jpg.
                Thanks for this screenshot. It helps to know how/where the data is stored.

                Comment

                Working...
                X