starred/photoprism

Fork 0

mirror of https://github.com/photoprism/photoprism.git synced 2026-03-02 22:57:18 -05:00

People: DeepStack go #588

New issue

Closed

opened 2026-02-19 23:11:45 -05:00 by deekerman · 15 comments

deekerman commented

2026-02-19 23:11:45 -05:00

Owner

Originally created by @ghost on GitHub (Dec 21, 2020).

Originally assigned to: @lastzero on GitHub.

I wish I knew how to use golang.. I came across DeepStackGo with is DeepStack for the golang

seems like it would make face identification simple. Can outline faces, add to the model and then build on finding the same taught ppl in other photos

Originally created by @ghost on GitHub (Dec 21, 2020). Originally assigned to: @lastzero on GitHub. I wish I knew how to use golang.. I came across DeepStackGo with is DeepStack for the golang seems like it would make face identification simple. Can outline faces, add to the model and then build on finding the same taught ppl in other photos

deekerman

2026-02-19 23:11:45 -05:00

closed this issue
added the
declined

idea
labels

deekerman commented

2026-02-19 23:11:45 -05:00

Author

Owner

@alexislefebvre commented on GitHub (Dec 22, 2020):

Could you please share a link to this library? I found one with this name but it looked unmaintained.

There already is an issue about face detection: #22

@alexislefebvre commented on GitHub (Dec 22, 2020): Could you please share a link to this library? I found one with this name but it looked unmaintained. There already is an issue about face detection: #22

deekerman commented

2026-02-19 23:11:46 -05:00

Author

Owner

@ghost commented on GitHub (Dec 22, 2020):

https://golang.hotexamples.com/examples/deepstack/DeepStack/Get/golang-deepstack-get-method-examples.html

https://github.com/johnolafenwa/deepstackgo

@ghost commented on GitHub (Dec 22, 2020): https://golang.hotexamples.com/examples/deepstack/DeepStack/Get/golang-deepstack-get-method-examples.html https://github.com/johnolafenwa/deepstackgo

deekerman commented

2026-02-19 23:11:46 -05:00

Author

Owner

@alexislefebvre commented on GitHub (Dec 22, 2020):

Thanks. Does it need to send photos to an external service to identify people?

@alexislefebvre commented on GitHub (Dec 22, 2020): Thanks. Does it need to send photos to an external service to identify people?

deekerman commented

2026-02-19 23:11:46 -05:00

Author

Owner

@ghost commented on GitHub (Dec 22, 2020):

No.. shouldn’t.. would use DeepStack locally

@ghost commented on GitHub (Dec 22, 2020): No.. shouldn’t.. would use DeepStack locally

deekerman commented

2026-02-19 23:11:46 -05:00

Author

Owner

@alexislefebvre commented on GitHub (Dec 22, 2020):

My bad, I misinterpreted the word “client” in the project description.

@alexislefebvre commented on GitHub (Dec 22, 2020): My bad, I misinterpreted the word “client” in the project description.

deekerman commented

2026-02-19 23:11:46 -05:00

Author

Owner

@graciousgrey commented on GitHub (Dec 23, 2020):

Thanks for investigating :) We will start implementing Face recognition in January!

@graciousgrey commented on GitHub (Dec 23, 2020): Thanks for investigating :) We will start implementing Face recognition in January!

deekerman commented

2026-02-19 23:11:46 -05:00

Author

Owner

@francescocarzaniga commented on GitHub (Jan 2, 2021):

@graciousgrey Offering DeepStack as an alternative to TensorFlow as well would be absolutely amazing. Tensorflow is a major pain in the butt to compile, and being able to connect Photoprism to a remote DeepStack server would be massive for me and, I believe, many others as well.

@francescocarzaniga commented on GitHub (Jan 2, 2021): @graciousgrey Offering DeepStack as an alternative to TensorFlow as well would be absolutely amazing. Tensorflow is a **major** pain in the butt to compile, and being able to connect Photoprism to a remote DeepStack server would be massive for me and, I believe, many others as well.

deekerman commented

2026-02-19 23:11:46 -05:00

Author

Owner

@jurgenhaas commented on GitHub (Feb 5, 2021):

I have integrated DeepStack with Home-Assistant earlier this week and I'm amazed how easy that was. It is so powerful, not only for face recognition but also for object and scene recognition. And as it runs locally with zero to none configuration effort, it would be much preferred to TensorFlow, because I don't want to use any of those services at all.

Their API is super slick and it should be very simple to integrate that into PhotoPrism. I'd be interested to help with this, if that's what you're looking for.

@jurgenhaas commented on GitHub (Feb 5, 2021): I have integrated DeepStack with Home-Assistant earlier this week and I'm amazed how easy that was. It is so powerful, not only for face recognition but also for object and scene recognition. And as it runs locally with zero to none configuration effort, it would be much preferred to TensorFlow, because I don't want to use any of those services at all. Their API is super slick and it should be very simple to integrate that into PhotoPrism. I'd be interested to help with this, if that's what you're looking for.

deekerman commented

2026-02-19 23:11:47 -05:00

Author

Owner

@jurgenhaas commented on GitHub (Feb 5, 2021):

Ah, here is a link to their docs: https://docs.deepstack.cc

@jurgenhaas commented on GitHub (Feb 5, 2021): Ah, here is a link to their docs: https://docs.deepstack.cc

deekerman commented

2026-02-19 23:11:47 -05:00

Author

Owner

@kn100 commented on GitHub (Feb 5, 2021):

It would be super ultra cool (and I am not necessarily requesting this, this would be a large undertaking!) if something like deepstack could run on my powerful desktop, which is not on all of the time, and when my desktop gets switched on Photoprism would see this and automatically send stuff over to be labelled. That way, you get past the fact Photoprism (which I want to be always available) is running on low power hardware, without having to have a high power image classification server running 24/7!

@kn100 commented on GitHub (Feb 5, 2021): It would be super ultra cool (and I am not necessarily requesting this, this would be a large undertaking!) if something like deepstack could run on my powerful desktop, which is not on all of the time, and when my desktop gets switched on Photoprism would see this and automatically send stuff over to be labelled. That way, you get past the fact Photoprism (which I want to be always available) is running on low power hardware, without having to have a high power image classification server running 24/7!

deekerman commented

2026-02-19 23:11:47 -05:00

Author

Owner

@robmarkcole commented on GitHub (Feb 21, 2021):

Deepstack can be run on a Jetson nano, which is low cost hardware. The combo of deepstack and photoprism would make a nice low cost & always on service

@robmarkcole commented on GitHub (Feb 21, 2021): Deepstack can be run on a [Jetson nano](https://developer.nvidia.com/embedded/jetson-nano-developer-kit), which is low cost hardware. The combo of deepstack and photoprism would make a nice low cost & always on service

deekerman commented

2026-02-19 23:11:47 -05:00

Author

Owner

@rahul003 commented on GitHub (Jun 29, 2021):

Is there any specific reason why Pigo was picked over Deepstack for this (https://github.com/photoprism/photoprism/issues/22)? I was wondering if I should look into this further for Photoprism

@rahul003 commented on GitHub (Jun 29, 2021): Is there any specific reason why Pigo was picked over Deepstack for this (https://github.com/photoprism/photoprism/issues/22)? I was wondering if I should look into this further for Photoprism

deekerman commented

2026-02-19 23:11:47 -05:00

Author

Owner

@lastzero commented on GitHub (Jun 30, 2021):

To reduce the number of dependencies. We would like to provide home users with a single binary instead of an enterprise micro-service architecture. Also we love the challenge. We may later add support for external services for advanced users.

@lastzero commented on GitHub (Jun 30, 2021): To reduce the number of dependencies. We would like to provide home users with a single binary instead of an enterprise micro-service architecture. Also we love the challenge. We may later add support for external services for advanced users.

deekerman commented

2026-02-19 23:11:47 -05:00

Author

Owner

@ohjables commented on GitHub (Aug 31, 2021):

Just a shout out for another service called double-take: https://github.com/jakowenko/double-take Claims to be a unified API that supports multiple detectors for face recognition/training/etc. Learned about it via frigate NVR but it might be cool to leverage the API in case things change in the future and you want to choose a different detector.

@ohjables commented on GitHub (Aug 31, 2021): Just a shout out for another service called double-take: https://github.com/jakowenko/double-take Claims to be a unified API that supports multiple detectors for face recognition/training/etc. Learned about it via frigate NVR but it might be cool to leverage the API in case things change in the future and you want to choose a different detector.

deekerman commented

2026-02-19 23:11:48 -05:00

Author

Owner

@lastzero commented on GitHub (Oct 4, 2024):

I am closing this issue as DeepStack does not seem to be maintained anymore:

https://github.com/johnolafenwa/DeepStack

Also note that we have started working on photoprism/photoprism-vision, which allows us to integrate additional AI/ML models through a REST API.

@lastzero commented on GitHub (Oct 4, 2024): I am closing this issue as DeepStack does not seem to be maintained anymore: - https://github.com/johnolafenwa/DeepStack Also note that we have started working on [`photoprism/photoprism-vision`](https://github.com/photoprism/photoprism-vision), which allows us to integrate additional AI/ML models through a REST API.