Insightface API

About

A simple deepinsight/insightface implementation with FastAPI for face verification.

It is recommended to test the available APIs from [GET] /docs

[GET] / - Root
- Check API status
[POST] /upload-selfie - Upload Selfie
- Upload selfie image file with person name and store it to database
[POST] /face-verification - Face Verification
- Upload selfie image file and a person name from the database to verify
[POST] /face-search - Face Search
- Upload selfie image file to search for similar faces in database
[PUT] /faces - Update Face
- Update face data (name or face data) in the database
[GET] /faces - Get Faces
- Get faces from database
[DELETE] /faces - Delete Face
- Delete a face from database using name from database
[POST] /analyze-image-url - Analyze Image Url
- Send image url to analyze (does not save to database)
[POST] /analyze-image-file - Analyze Image File
- Upload image file to analyze (does not save to database)
[POST] /compute-selfie-image-files-similarity - Compute Selfie Image Files Similarity
- Compute similarity of 2 selfie image files (does not save to database)

First you must run setup_model.sh to load the model.

Using docker-compose (RECOMMENDED):

docker-compose up

Using pipenv (you'll need to setup postgresql and .env on your own):

pipenv run uvicorn app.main:app --reload

Using poetry (you'll need to setup postgresql and .env on your own):

poetry run uvicorn app.main:app --reload

You can download the model files and move them into ~/.insightface/models manually or just use setup_model.sh

Original source:

Alternative:

Name		Name	Last commit message	Last commit date
Latest commit History 82 Commits
app		app
dockerfiles		dockerfiles
.dockerignore		.dockerignore
.env.example		.env.example
.gitignore		.gitignore
Pipfile		Pipfile
Pipfile.lock		Pipfile.lock
README.md		README.md
docker-compose.yml		docker-compose.yml
init_model.py		init_model.py
poetry.lock		poetry.lock
pyproject.toml		pyproject.toml
setup_model.sh		setup_model.sh