Data

Tools

Indexes

Applications

Experiments

Open Source Science

fplnet

FPLNet : Face Parsing and Landmarks with Cascaded ConvNets

https://github.com/msch8791/fplnet

Science Score: 57.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ✓
    CITATION.cff file
    Found CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ✓
    DOI references
    Found 1 DOI reference(s) in README
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (13.1%) to scientific vocabulary
Last synced: 10 months ago · JSON representation ·

Repository

FPLNet : Face Parsing and Landmarks with Cascaded ConvNets

Basic Info
  • Host: GitHub
  • Owner: MSch8791
  • License: bsd-3-clause
  • Language: Python
  • Default Branch: main
  • Homepage:
  • Size: 819 KB
Statistics
  • Stars: 0
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 2
Created about 2 years ago · Last pushed almost 2 years ago
Metadata Files
Readme License Citation

README.md

FPLNet

FPLNet : Face Parsing and Landmarks with Cascaded ConvNets

demo demo

Summary

This repository contains the FPLNet models, a facial landmarks and face parsing estimator. It also contains source code to run the inference on the model and provide a test tool to showcase its abilities.

How it works

The processing pipeline consists of two principal steps : 1. Detect faces in an image or video frame 2. Infer with FPLNet the landmarks and face parsing segmentation for each detected face

The models can be easily deployed in third-party projects by adding a few lines of code. Check the test.py file for an example.

Getting started

Installing

  1. Create a virtual environment (python >= 3.9).

  2. Clone the repository bash git clone https://github.com/MSch8791/FPLNet.git

  3. Install the dependencies with pip bash pip install -r requirements.txt

  4. The trained models are provided in the models directory. Download them with Git LFS bash git lfs pull

  5. Test using the test script bash cd src/ python test.py --image=../test/test_image.jpg --output=../test/test_output.jpg

Performances

Model name | Test dataset | Landmarks NMEinter-eye | Landmarks RMSE | Parsing F1-score | :---: | :---: | :---: | :---: | :---: | fplnet256LaPa4c20240517 | LaPa test | 0.02132 (2.13%) | 2.5659 | TO DO |

BibTeX Citation

Please consider citing this project in your publications if it helps your research or projects. BibTeX reference is as follows. @misc{Scherer_FPLNet_Face_2024, author = {Scherer, Michael}, title = {FPLNet : Face Parsing and Landmarks with Cascaded ConvNets}, month = {January}, year = {2024}, publisher = {GitHub}, journal = {GitHub repository}, howpublished = {\url{https://github.com/MSch8791/FPLNet}} doi = {10.5281/zenodo.13821451} }

Acknowledgments

The FPLNet model has been trained on the LaPa train dataset available here : https://github.com/jd-opensource/lapa-dataset

The face detector model used in the demo/test script is YuNet through OpenCV.

Owner

  • Login: MSch8791
  • Kind: user

Citation (CITATION.cff) 

cff-version: 1.2.0
message: "Please consider citing this project in your publications if it helps your research or projects. Cite it as below."
authors:
  - family-names: "Scherer"
    given-names: "Michael"
title: "FPLNet : Face Parsing and Landmarks with Cascaded ConvNets"
version: 1.0.0
doi: 10.5281/zenodo.13821451
date-released: 2024-05-16
preferred-citation:
  type: generic
  authors:
  - family-names: "Scherer"
    given-names: "Michael"
  title: "FPLNet : Face Parsing and Landmarks with Cascaded ConvNets"
  year: 2024
  doi: 10.5281/zenodo.13821451

GitHub Events

Total
Last Year

Dependencies

requirements.txt pypi
  • numpy >=1.23.5
  • onnxruntime >=1.15.1
  • opencv-python >=4.9.0.80