This project is about Face 'Liveness' and Anti-Spoofing' detection.
It is Not about simple face detection.
Payment for work done. Plus good Bonus for a fast delivered end result.
I am looking for someone with current experience to create an Anti-Spoofing / Liveness Face Recognition Library that can be used in Android, iPhone and via a Web page at the server side. The Web page component will likely be JavaScript for the user-interface and run back to the logic processing at the server end.
I have seen and tried a few GitHub projects that if combined might create the result I am looking for.
OBJECTIVES
A) Create using some GitHub projects, a Web process for Face Detection and Matching (Recognition)
Can use existing OpenCV and Eigenface code for Matching/Recognition
B) Create a similar process in a Library that can be used in Android and iOS
A-1. Easy
Combine a few GitHub projects for Face using JavaScript / Node.js etc.
A-3. Difficult
Anti-Spoofing. Using one or combination of GitHub projects combat against Spoofing that uses printed images, photos or phones images / video.
By Anti-Spoofing I mean that the code will be able to detect if the face is Human or a Photo representation.
PERSON
Ideally someone who knows about Models, Alignment, CNN / Caffe, Liveness, AntiSpoofing
LANGUAGES AND LIBRARIES
C++ OpenCV
C# Eigenface
Python Face Recognition
JavaScript
JAVA
Android
This project can be tackled in a number of ways and I am open to suggestions on how that can be done.
For example ;
I) Capture the Face points to determine it is a Face (Detection easy)
ii) Request the user moves their face / smiles / closes an eye (Liveness easy)
iii) Apply an Anti-Spoofing algorithm to the image using a trained model (this is the tough bit)
There is plenty of existing code for I) and ii) for 64 point and 106 point which is better and I have also tried them for Detection with success.
ANTI-SPOOFING
Below are a number of GitHub projects some of which look like they might do the job.
Some are in Chinese so you might need Google Translate. I did and it is fine.
Some are in Python so for a Phone solution they will need to be run inside something like 'Chaquopy' or converting it to JavaScript or C# or Java etc.
There are quite a few Python code converters available even online ones. If I need to purchase one for you I can probably do that.
https://github.com/SoftwareGift/Face-Anti-spoofing-for-RGBPROCESSING RESPONSE TIME
This is an important consideration and this matter should be taken into account.
B-1.
LIBRARY
The Android and iPhone libraries will be whatever is most appropriate for those platforms.
You will need to be able to ;
a) create the Library
b) create a simple test app for each platform to demonstrate it working
Reply with the word BUTTER so I know you have read the job.
About the recuiterMember since Mar 14, 2020 Raju M
from Georgia, United States