CompreFace VS insightface

# https://github.com/exadel-inc/CompreFace/blob/master/docs/Face-services-and-plugins.md)

I looked at Double Take (UI that lets you do the training of your face recondition easily) and found CompreFace as one of models they support. It looks like what I need but there is a catch... no OpenVino (intel CPUs AI accelerator) support. I really like my low power NVR setup and would like keep it that way. Running AI on CPU without acceleration is both power inefficient and much slower. I have a spare low end GPU but if dump it in the system the current AI acceleration brakes... (I know I can prob fix it but that is a rabbit hole I would prefer to avoid).

This has good UI and also good recognition: https://github.com/exadel-inc/CompreFace

wget -q -O tmp.zip 'https://github.com/exadel-inc/CompreFace/releases/download/v1.1.0/CompreFace_1.1.0.zip' && unzip tmp.zip && rm tmp.zip

https://github.com/exadel-inc/CompreFace/ Is already a good locally-run API for face detection and facial recognition. Thankyou for your suggestion.

If you are looking for open source stuff and are able to self-host it, maybe have a look at CompreFace.

As almost always, on GitHub https://github.com/exadel-inc/CompreFace

Looking into this I found Compreface (https://exadel.com/solutions/compreface/) an open source face recognition software. There are alread some controb scripts, like contrib/photils.lua, who take some images, run them through a tool, then tag them with data coming from the tool. Converting this to use Compreface looks likea promising avenue.

I haven't used it for search, but I believe Insightface's embeddings can be used for this purpose. https://insightface.ai/

This is Picsi.ai/InsightFace Discord server: https://discord.gg/Ym3X8U59ZN The COO of the company, enforcing these strikes is Discord user unmoved.mover This is the github of insightface: https://github.com/deepinsight/insightface

Insightface offering improved model but closed source via discord bot.

https://github.com/deepinsight/insightface/issues/2294

https://github.com/deepinsight/insightface/issues/2315

The easiest I've found so far is to use InsightFace. When you call it on an image, it will return both a 2D and 3D mesh showing the orientation of the face.

Let's say I want to make a flask api using insightface on a linode server. How much ram do I really need?

I have tried Insightface ( https://insightface.ai/ ). But it struggles with face edges. i have used GFPGAN too, but that still leaves blurs around the face. Notice below https://www.youtube.com/watch?v=jWIAVjVeG1I

Mainly as the base model. https://insightface.ai/ There was some post processing done to further improve quality. https://github.com/TencentARC/GFPGAN

in this page says that you need the onxruntime to run it https://github.com/deepinsight/insightface/tree/master/python-package

According to this, they're working on a paper for the 256 model right now.

This is a tongue-in-cheek variation of a WIP for Mat Cauthon. RJ has a list of casting choices and a young James Garner was his selection for Mat Cauthon. The illustration itself is from my own WIP, and no, Mat doesn't look like James Garner in my version. I used Insight Faceswap, https://insightface.ai/ along with a pic of a young James Garner. Made me laugh...