Activity

Pegoku

I’ve updated the outline and selection logic so it now only shows the most significative outline

Attachment
0
Pegoku

Bulk Classify works!
I was able to get the right-click to work, and I can now classify multiple faces at once!

Attachment
0
Pegoku

I got the classifyFace button working! With it you can classify faces and assign them to a person!
PS: it also autoreloads when you do so

Attachment
0
Pegoku

I’ve been working on getting the face select working, now, you can select the person you want to view, and it will show the images assigned to it

Attachment
Attachment
0
Pegoku

I’ve been working on the header.
First, I wanted to do it by hand, ie, without using any lib, but then I realized it would be easier and better, and way more consistent if I just used an already made component set (eg. Radix, the same one Frigate uses).
Also got the header to display the correct info

Attachment
0
Pegoku

I fixed the grouping and layout, it now is quite nicer.
I also added titles on top with the most frequent face, now I will add the avg confidence %.
It is getting quite close to the original UI!

Attachment
0
Pegoku

oops, I think I messed up a bit the code trying to group the faces, lol

Attachment
0
Pegoku

I’ve been working on the page, The hardest part for now, was to be able to get the images pulled from the Frigate server, as I needed to auth (something I should have expected). To solve that problem, I created an API where you indicate the name and face file, and you’ll get the image.
I also was able to parse the identified names, and %.

PS: blurred the images and names for privacy.

Attachment
0
Pegoku

To start, I did a simple script where it logins into Frigate using a token or user/pass and then it gets all the faces from it.
This is simple, but I think it’s a good start. I will now start a nextjs project and create a simple gui where it displays the faces with names, %s, etc.
I’ll try to do a similar layout to the real frigate (for now)

Attachment
0