lineup

recognition of faces for the web

Demo

Why

Puzzle pieces ask us to recognise from reference:

graph LR;
    A(((10001/10001 pieces left))) --> B([Blue Puzzle Piece])
    B -.-> C(Sky?)
    B -.-> D(Sea?)
    B -.-> E(Blue petticoat?)

Loading

graph LR;
    A(((11 pieces left))) --> B([Green Puzzle Piece])
    B --> E(Tree on Left)
    A ==> C([Dark-Blue Puzzle Piece])
    C --> D(Sea)
    A ==> F([Light-Blue Puzzle Piece])
    F --> G(Sky)

Loading

To reduce the Δtime it takes to build up references, machine learning libraries can be used.

As PJ Vogt's Search Engine podcast discusses, this is possible for faces.

Mechanism

Using face-recognition to build up the reference list in a database, the lineup browser extension can allow for all faces found while browsing to later be checked.

This software uncouples the browsing and the reference list.

Multiple people, on different browsers, can share the same database of references.

graph LR;
    A(((face_db))) <--> B([face])
    B <--> E(firefox)
    A <--> F([face])
    F <--> G{{librewolf}}

Loading

graph BT;
    B(firefox) <--> A[(face_db)]
    C{{librewolf}} <--> A
    D(firefox) <--> A
    E(firefox) <--> A
    F{{librewolf}} <--> A

Loading

Installation Requirements & Usage

Requirements - Python & Database

redacted due to EU AI Act

Installation - Client Side

redacted due to EU AI Act

Usage

1. launch elasticsearch instance, wait until the output about 'health'
2. launch $repo/lineup_svr/app.py using python3 app.py
3. launch firefox, go to about:debugging
4. select This Browser
5. select load temporary extension and choose $repo/lineup_ext/manifest.json

Licensing

Open Source License

This software is available under the MIT License.

Commercial use

redacted due to EU AI Act