I am unable to run facial recognition without Digikam crashing. Is this a common problem, and are there known issues or workarounds for this? I’m running Digikam 7.1 on Windows 10, with a database of about 100,000 pictures. All databases are set to MySQL Server (experimental), which connects to a Synology NAS running MariaDB 10.3.21 over the network. This is a bit slow, but my wife and I both access the database from separate PCs and need to share tags. To run the recognition, I go to the People tab, choose Workflow -> Detect faces, Search In -> Whole albums collection. When this completed, I expected to manually tag a variety of faces in the Unknown tag and then repeat the process with Workflow -> Recognize faces. I assume this would reclassify most of the pictures from Unknown to the proper name tag, except for those that end up in Unconfirmed and need manual tagging. Instead, Digikam adds about 800 faces to Unknown and then crashes (see the Windows EventVwr Application Error below). When I repeat the process with Recognize Faces, very few faces get recognized and I find myself mostly manually tagging. Faces are primarily friends and family over a 20 year period, so I have tagged people at points of life from babies through late teens. Questions:
Thank you! (App crash log is below) Log Name: Application Source: Application Error Date: 10/9/2020 4:21:01 PM Event ID: 1000 Task Category: (100) Level: Error Keywords: Classic User: N/A Computer: PC-VM Description: Faulting application name: digikam.exe, version: 0.0.0.0, time stamp: 0x00000000 Faulting module name: libstdc++-6.dll, version: 0.0.0.0, time stamp: 0x00000000 Exception code: 0x40000015 Fault offset: 0x000000000002070a Faulting process id: 0x2ddc Faulting application start time: 0x01d69c00905c92da Faulting application path: C:\Program Files\digiKam\digikam.exe Faulting module path: C:\Program Files\digiKam\libstdc++-6.dll Report Id: 762bb3d4-ed93-4e35-8912-ab928529a5c2 Faulting package full name: Faulting package-relative application ID: |
Well, we already have various bug reports on this. We're working on it. I
understand correctly that face detection works, it crashes in face recognition? Try the digiKam-7.2.0-beta1 from here: https://files.kde.org/digikam/ Maik Am Samstag, 10. Oktober 2020, 18:58:28 CEST schrieb [hidden email]: > I am unable to run facial recognition without Digikam crashing. Is this a > common problem, and are there known issues or workarounds for this? > > > > I'm running Digikam 7.1 on Windows 10, with a database of about 100,000 > pictures. All databases are set to MySQL Server (experimental), which > connects to a Synology NAS running MariaDB 10.3.21 over the network. This is > a bit slow, but my wife and I both access the database from separate PCs > and need to share tags. > > > > To run the recognition, I go to the People tab, choose Workflow -> Detect > faces, Search In -> Whole albums collection. When this completed, I expected > to manually tag a variety of faces in the Unknown tag and then repeat the > process with Workflow -> Recognize faces. I assume this would reclassify > most of the pictures from Unknown to the proper name tag, except for those > that end up in Unconfirmed and need manual tagging. > > > > Instead, Digikam adds about 800 faces to Unknown and then crashes (see the > Windows EventVwr Application Error below). When I repeat the process with > Recognize Faces, very few faces get recognized and I find myself mostly > manually tagging. Faces are primarily friends and family over a 20 year > period, so I have tagged people at points of life from babies through late > teens. > > > > Questions: > > 1. How can I stop the crashing? Is there a better log I could upload? > 2. Is my workflow (Detect faces first, then Recognize Faces second) > correct? > 3. Am I correct in using varying ages for training data (the initial > manual tags), or should I only pick one age and let the AI sort it out? > > > > Thank you! (App crash log is below) > > > > > > Log Name: Application > > Source: Application Error > > Date: 10/9/2020 4:21:01 PM > > Event ID: 1000 > > Task Category: (100) > > Level: Error > > Keywords: Classic > > User: N/A > > Computer: PC-VM > > Description: > > Faulting application name: digikam.exe, version: 0.0.0.0, time stamp: > 0x00000000 > > Faulting module name: libstdc++-6.dll, version: 0.0.0.0, time stamp: > 0x00000000 > > Exception code: 0x40000015 > > Fault offset: 0x000000000002070a > > Faulting process id: 0x2ddc > > Faulting application start time: 0x01d69c00905c92da > > Faulting application path: C:\Program Files\digiKam\digikam.exe > > Faulting module path: C:\Program Files\digiKam\libstdc++-6.dll > > Report Id: 762bb3d4-ed93-4e35-8912-ab928529a5c2 > > Faulting package full name: > > Faulting package-relative application ID: |
DigiKam-7.2.0-Beta1 fixes problems related to MySQL and the face database.
First you should then rebuild the training data. Run Maintenance Tool-> Face Recognition section-> [x] Rebuild Training Database. Maik Am Samstag, 10. Oktober 2020, 19:07:15 CEST schrieben Sie: > Well, we already have various bug reports on this. We're working on it. I > understand correctly that face detection works, it crashes in face > recognition? Try the digiKam-7.2.0-beta1 from here: > > https://files.kde.org/digikam/ > > Maik > > Am Samstag, 10. Oktober 2020, 18:58:28 CEST schrieb [hidden email]: > > I am unable to run facial recognition without Digikam crashing. Is this a > > common problem, and are there known issues or workarounds for this? > > > > > > > > I'm running Digikam 7.1 on Windows 10, with a database of about 100,000 > > pictures. All databases are set to MySQL Server (experimental), which > > connects to a Synology NAS running MariaDB 10.3.21 over the network. This > > is a bit slow, but my wife and I both access the database from separate > > PCs and need to share tags. > > > > > > > > To run the recognition, I go to the People tab, choose Workflow -> Detect > > faces, Search In -> Whole albums collection. When this completed, I > > expected to manually tag a variety of faces in the Unknown tag and then > > repeat the process with Workflow -> Recognize faces. I assume this would > > reclassify most of the pictures from Unknown to the proper name tag, > > except for those that end up in Unconfirmed and need manual tagging. > > > > > > > > Instead, Digikam adds about 800 faces to Unknown and then crashes (see the > > Windows EventVwr Application Error below). When I repeat the process with > > Recognize Faces, very few faces get recognized and I find myself mostly > > manually tagging. Faces are primarily friends and family over a 20 year > > period, so I have tagged people at points of life from babies through late > > teens. > > > > > > > > Questions: > > > > 1. How can I stop the crashing? Is there a better log I could upload? > > 2. Is my workflow (Detect faces first, then Recognize Faces second) > > correct? > > 3. Am I correct in using varying ages for training data (the initial > > manual tags), or should I only pick one age and let the AI sort it out? > > > > > > > > Thank you! (App crash log is below) > > > > > > > > > > > > Log Name: Application > > > > Source: Application Error > > > > Date: 10/9/2020 4:21:01 PM > > > > Event ID: 1000 > > > > Task Category: (100) > > > > Level: Error > > > > Keywords: Classic > > > > User: N/A > > > > Computer: PC-VM > > > > Description: > > > > Faulting application name: digikam.exe, version: 0.0.0.0, time stamp: > > 0x00000000 > > > > Faulting module name: libstdc++-6.dll, version: 0.0.0.0, time stamp: > > 0x00000000 > > > > Exception code: 0x40000015 > > > > Fault offset: 0x000000000002070a > > > > Faulting process id: 0x2ddc > > > > Faulting application start time: 0x01d69c00905c92da > > > > Faulting application path: C:\Program Files\digiKam\digikam.exe > > > > Faulting module path: C:\Program Files\digiKam\libstdc++-6.dll > > > > Report Id: 762bb3d4-ed93-4e35-8912-ab928529a5c2 > > > > Faulting package full name: > > > > Faulting package-relative application ID: |
In reply to this post by Maik Qualmann
Maik, thank you for the response, this seems to have resolved the issue. I downloaded 7.2-beta1, rebuilt the training database, and re-ran the “Detect faces” process. It completed successfully with no crashes. I am now running the “Recognize faces” process, and it’s at 4% but looks like it’s working well. One comment: when Digikam completed detecting faces, the “Unknown” People tag displayed at (0). During detection, this had been up over 50,000. Once I selected “Recognize Faces” and clicked “Scan collection for faces”, this counter reset to the correct number (over 59,000), and the counter began decrementing as those faces got recognized. Is “Unknown” being set to (0) a bug, or by design? If a bug, should I report it? Thank you again for the assistance. > Date: Sat, 10 Oct 2020 19:27:28 +0200 > From: Maik Qualmann <[hidden email]> > To: digiKam - Home Manage your photographs as a professional with the > power of open source <[hidden email]> > Subject: Re: [digiKam-users] Crash during facial recognition > > DigiKam-7.2.0-Beta1 fixes problems related to MySQL and the face database. > First you should then rebuild the training data. Run Maintenance Tool-> Face Recognition section-> [x] Rebuild Training Database. > > Maik > > Am Samstag, 10. Oktober 2020, 19:07:15 CEST schrieben Sie: > > Well, we already have various bug reports on this. We're working on > > it. I understand correctly that face detection works, it crashes in > > face recognition? Try the digiKam-7.2.0-beta1 from here: > > > > https://files.kde.org/digikam/ > > > > Maik > > > > > Am Samstag, 10. Oktober 2020, 18:58:28 CEST schrieb [hidden email]: > > > I am unable to run facial recognition without Digikam crashing. Is > > > this a common problem, and are there known issues or workarounds for this? |
The "Unknown" tag is a bit virtual, if you haven't assigned it to an image, it
will be displayed as (0) in the tag tree. In the people view you can see the "virtual" number of faces that are "Unknown". Maybe a little strange, but normal. Maik Am Montag, 12. Oktober 2020, 10:00:03 CEST schrieb [hidden email]: > Maik, thank you for the response, this seems to have resolved the issue. I > downloaded 7.2-beta1, rebuilt the training database, and re-ran the "Detect > faces" process. It completed successfully with no crashes. I am now running > the "Recognize faces" process, and it's at 4% but looks like it's working > well. > > > > One comment: when Digikam completed detecting faces, the "Unknown" People > tag displayed at (0). During detection, this had been up over 50,000. > > > > Once I selected "Recognize Faces" and clicked "Scan collection for faces", > this counter reset to the correct number (over 59,000), and the counter > began decrementing as those faces got recognized. > > > > Is "Unknown" being set to (0) a bug, or by design? If a bug, should I report > it? > > > > Thank you again for the assistance. > > > Date: Sat, 10 Oct 2020 19:27:28 +0200 > > > > From: Maik Qualmann <[hidden email] <mailto:[hidden email]> > > > > > > To: digiKam - Home Manage your photographs as a professional with the > > > > power of open source <[hidden email] > > <mailto:[hidden email]> > > > > Subject: Re: [digiKam-users] Crash during facial recognition > > > > > > > > DigiKam-7.2.0-Beta1 fixes problems related to MySQL and the face database. > > > > > > First you should then rebuild the training data. Run Maintenance Tool-> > > Face Recognition section-> [x] Rebuild Training Database. > > > Maik > > > > Am Samstag, 10. Oktober 2020, 19:07:15 CEST schrieben Sie: > > > Well, we already have various bug reports on this. We're working on > > > > > > it. I understand correctly that face detection works, it crashes in > > > > > > face recognition? Try the digiKam-7.2.0-beta1 from here: > > > > > > > > > > > > https://files.kde.org/digikam/ > > > > > > > > > > > > Maik > > > > > > > Am Samstag, 10. Oktober 2020, 18:58:28 CEST schrieb [hidden email] > > <mailto:[hidden email]> : > > > > I am unable to run facial recognition without Digikam crashing. Is > > > > > > > > this a common problem, and are there known issues or workarounds for > > this? |
Free forum by Nabble | Edit this page |