Question 1

Does Photonix run on Raspberry Pi / ARM?

Accepted Answer

Yes, this feature was added in April 2021 and we have tested on Raspberry Pi 4 Model B and Pi 2 Model B. This covers both major architectures: ARM32v7 and ARM64v8.
ARM Single Board Computers (SBCs) are ideal for low-powered home servers and allows you to access your photos while you are away from home.
Due to the fact that most ARM computers are lower power and have less memory and CPU performance than a typical computer there are a few things we recommend to ensure it doesn't get overloaded.

Turn off ML analysis to start with. You are given the option to turn off things like object detection, style detection, face recognition when you go through the onboarding setup. You can turn these back on from the settings screen later.
Start by loading in a small portion of your photo collection. Once you you have a sense of how long it takes to import you can decide whether to continue with the rest.

You can read more in our announcement blog post.

Question 2

Is it safe to use Photonix with my existing folder structure?

Accepted Answer

Photonix doesn’t make any changes to your existing photos – it only generates tags and other data in it’s own  database. If you upload photos through Photonix’s web or mobile interfaces then Photonix will add the uploaded files to your photos folder but they will remain unmodified.
If you want to be extra certain that Photonix can’t modify your photos you can bind the folder as a read-only Docker volume using something like -v volume-name:/path/in/container:ro in Docker run or adding :ro to the end of a volume path in your docker-compose.yml file.

Question 3

How would this photo organizer fit into my existing process?

Accepted Answer

Well, you can think of your photos as going through a pipeline. First you are going to be out and about with your digital camera actually taking the photos. After shooting you’ll typically download the files onto your computer and do your editing. It’s after this editing stage where the waters can get a bit murky and there is no obvious path to take. You’ll have some more files now and probably some more folders but where should they live for the long term?
We’ve tried to make this photo organizing software adapt to you so you don’t have to change for it. Some users will have an external hard drive or two for their files because they either won’t fit on their computer’s hard drive or as a backup. Other more advanced users may have a home server on their network so that they can share files between computers in their home.
As long as Photonix knows which folders new files (RAW and processed) will appear in, the photo organizer component will get to work, building its database. It can use the files directly if you are running on a single computer and need to save storage space or it can copy the files if running in a server/cloud scenario. Detected photos should show nearly instantly in Photonix’s user interface. The more advanced AI analysis will happen in the background (as it can take a bit of time) and then you’ll be able to use all the more advanced forms of search.

Question 4

Does Photonix support videos?

Accepted Answer

Not yet, unfortunately.
We aim to add limited support for video files soon but right now, we’re trying to do the best job we can for photographers.

Question 5

Can Photonix help with organizing the RAW files that I use for editing?

Accepted Answer

Yes, absolutely. We’re glad you asked! Photonix is photo management software for professionals, and we know this is an area that isn’t catered for with most services like Google Photos, Flickr, 500px and desktop applications such as Picasa, DigiKam and Shotwell.
When Photonix imports your photos, it doesn’t just look at the JPEGs but reads all the main manufacturer’s RAW file formats (.CR2, .NEF, .DNG, .ARW, .RWL and more). If you have a RAW format that we can’t open then please let us know and we’ll try to support it.
Don’t worry about the RAW files showing up as duplicates though — our de-duplication algorithm is usually smart enough to realise multiple files are actually the same photo — just a different "version". We try to combine duplicate photos whether they’re JPEGs produced by the camera, RAW files or edits exported from Lightroom as TIFF.

Question 6

Can I start using it right now?

Accepted Answer

The installation process requires quite a bit of technical know-how currently, but we are working on making it really simple. The source code is open source and other developers have been busily testing it out on their own images. By the time we reach our 1.0 release it will be ready for the general public to use. We’re building this free photo organizer in the open to produce the best solution through collaboration.
If you are technically inclined then please take a look at our Installation Instructions. We'd really appreciate your feedback.
Please add your email address to the mailing list form below and we’ll let you know when ready it is for the general public.

Question 7

Will Photonix work with my other software tools?

Accepted Answer

We understand that keen photographers have their go-to editing tools and may want to experiment with other pieces of software from time to time. For this reason, we don’t require you to install plugins to integrate with us. You just need to tell Photonix which folders on your computers you are going to export edits to and it will continually watch it for changes in the background. This works with local folders, network mounted filesystems and external hard drives.
Because of this, it doesn’t matter whether you’re an Adobe Lightroom loyalist, a big Darktable fan, a Photoshop guru or prefer something else.

Question 8

Does the face recognition model respect privacy?

Accepted Answer

No personal data (images, metadata or AI learnings) ever leaves your local server if your run your own instance of Photonix.
We were able to develop face recognition without sending any data to the cloud and break it down so the only part that needs retraining specifically to your collection of faces is very fast. You can read the in-depth steps of our face detection and recognition model here if you are so inclined.
All the computationally intensive steps mentioned in that document are performed by pre-trained, industry-standard models. This gets us to the point where we have a fingerprint (or embedding) array of 128 numbers to roughly represent each detected face.
The very last step labelled Similarity Calculation / Clustering / Classification is the only part of the model that is built on the local machine. This similarity index is very quick to build (typically a few seconds) and just means that we don't have to compare the current face embedding individually with every other face embedding already in the system when we want to find the closest match.

Question 9

Does it create web-friendly JPEG images for fast loading?

Accepted Answer

Yes you should find that images in the UI are fast-loading JPEGs. Photonix has a thumbnailing service built in which produces smaller versions of photos for display. You won't tend to see the original files unless you click the download icon. The most frequently seen JPEGs are:

256x256px, 'cover', 50% quality, ~5-10KB
3840x3840px, 'contain', 75% quality, ~500KB-1MB

There is naturally a trade-off between size and quality. On our photo detail screen we load a slightly larger images than you may expect but this is because we still want to show a reasonable level of detail if the user tries to zoom in.
With relation to raw files, a full-size JPEG is created of all of these which is then used to create smaller thumbnails from.

Question 10

Does Photonix do face recognition and labelling of people?

Accepted Answer

Yes, we have support for detecting, grouping and labelling faces. This means that if you label some of your friends, others will get automatically labelled so you can quickly find other photos of them.
These are the main steps of the process for each photo:

Detect face positions and sizes in photos that are imported.
Crop, transform and align face to correct for people positioning their heads differently.
Compute embedding “fingerprint” for each face image.
Build a Nearest Neighbours search index to quickly compute distance between face embeddings. If two faces are close enough together, tag them as being the same person.
Provide user interface so a human can label, accept or reject the face groups that were found.

More details are given in our announcement blog post and Image Analysis documentation.

FAQs

Does Photonix run on Raspberry Pi / ARM?

Is it safe to use Photonix with my existing folder structure?

How would this photo organizer fit into my existing process?

Does Photonix support videos?

Can Photonix help with organizing the RAW files that I use for editing?

Can I start using it right now?

Will Photonix work with my other software tools?

Does the face recognition model respect privacy?

Does it create web-friendly JPEG images for fast loading?

Does Photonix do face recognition and labelling of people?

Ironicbadger

Jun 2021

BMHz

Apr 2021

Thors_Son

Jun 2021