Hello, I am learning about the autotagging feature/services available for IMatch, which I would like to use. Before I start to use one of them, or read through 20 pages of legalese on these services sites, I'd like to know if there is any consensus or opinions on the privacy these services afford to users. i.e. is one preferred for privacy? Do some guarantee to delete the thumbnails that are uploaded to use the services? I've done some googling but at least so based on the search terms I'm using so far I just get a bunch of irrelevant hits on privacy concerns around Facbook photos etc.
Thanks for any advice.
Each service provides contact addresses via a web form or email address. This probably the best way to ask specific questions about the service or license agreement / privacy.
Like, how long they retain images you have used with their service, if they store the actual image or only a hash etc.
Except for imagga, which is based on the EU and implements the GDPR (https://imagga.com/privacy), all other services are based in the U.S. and thus have the standard privacy issues caused by U.S. laws and regulations.
If you're concerned about privacy, you should have a look at Excire. It writes XMP metadata into the photos and it runs locally on your PC - like IMatch for face detection.
Excire is definitely something to check out.
I have evaluated Excire technology while doing AutoTagger research in early versions of IMatch 2020 Beta.
I've signed an NDA so I cannot comment.
In the end, I've decided to go with cloud-based SaaS providers for AutoTagger.
It made it easier to offer IMatch users multiple services/choices and to give them access to providers like Clarifai which offer different AI models for specific purposes like wedding or food photography, imagga which is based in Europe and has to apply the GDPR. Google's AI is including the intelligence and learnings from Google search, which allows it to identify buildings and hot spots often visited by tourists - which produces superior results for typical travel and vacation photography. Unfortunately, all U.S. based companies cannot provide proper privacy, due to U.S. laws and regulations. From a non-US / European perspective.
That being said, the architecture of AutoTagger is very open and flexible and I can easily integrate other SaaS (cloud-base) services easily, if there is demand.
That being said, IMatch offers many tools and features to quickly classify and tag (keyword) large volumes of images without the help of any external service.
AutoTagger is definitely very helpful when you start to manage very large collections with tens or even hundreds of thousands of files. Spending a hundred bucks to get 100,000 files tagged is nothing compared to the manual work required to do the same. Corporate and institutional users of IMatch benefit a lot from that.
If you only have to deal with "personal" photos taken during vacations, parties, travels, festivals etc., you can most likely work very efficiently with some Metadata Templates and the Keywords Panel. No need to use 3rd party AI services at all.