Our facial recognition algorithm is globally top-ranked by NIST in the FRVT 1:1 leaderboards.

Latest NIST FRVT evaluation report 2024-12-20

🆔 ID Document Liveness Detection - Linux - Here

🤗 Hugging Face - Here

📚 Product & Resources - Here

🛟 Help Center - Here

💼 KYC Verification Demo - Here

🙋‍♀️ Docker Hub - Here

sudo docker pull kbyai/face-recognition:latest
sudo docker run -e LICENSE="xxxxx" -p 8081:8080 -p 9001:9000 kbyai/face-recognition:latest

FaceRecognition-Docker

Overview

This repository demonstrates an advanced face recognition technology by implementing face comparison based on face feature extraction and face matching algorithm, which was implemented via a Dockerized Flask API.
It includes features that allow for testing face recognition between two images using both image files and base64-encoded images.

In this repo, we integrated KBY-AI's face recognition solution into Linux Server SDK by docker container.
We can customize the SDK to align with customer's specific requirements.

◾FaceSDK(Server) Details

Face Liveness Detection	🔽 Face Recognition
Face Detection	Face Detection
Face Liveness Detection	Face Recognition(Face Matching or Face Comparison)
Pose Estimation	Pose Estimation
68 points Face Landmark Detection	68 points Face Landmark Detection
Face Quality Calculation	Face Occlusion Detection
Face Occlusion Detection	Face Occlusion Detection
Eye Closure Detection	Eye Closure Detection
Mouth Opening Check	Mouth Opening Check

◾FaceSDK(Server) Product List

No.	Repository	SDK Details
1	Face Liveness Detection - Linux	Face Livness Detection
2	Face Liveness Detection - Windows	Face Livness Detection
3	Face Liveness Detection - C#	Face Livness Detection
➡️	Face Recognition - Linux	Face Recognition
5	Face Recognition - Windows	Face Recognition
6	Face Recognition - C#	Face Recognition

To get Face SDK(mobile), please visit products here:

Try the API

Online Demo

You can test the SDK using images from the following URL: https://web.kby-ai.com

Postman

To test the API, you can use Postman. Here are the endpoints for testing:

• Test with an image file: Send a POST request to http://18.221.33.238:8081/compare_face.

• Test with a base64-encoded image: Send a POST request to http://18.221.33.238:8081/compare_face_base64.

  You can download the Postman collection to easily access and use these endpoints. click here

  

SDK License

This project uses KBY-AI's Face Recognition Server SDK, which requires a license per machine.

• The code below shows how to use the license:

  FaceRecognition-Docker/app.py

  Lines 26 to 36 in 5c6bdaf

  	machineCode = getMachineCode()
  	print("machineCode: ", machineCode.decode('utf-8'))
  	try:
  	with open(licensePath, 'r') as file:
  	license = file.read()
  	except IOError as exc:
  	print("failed to open license.txt: ", exc.errno)
  	print("license: ", license)
  	ret = setActivation(license.encode('utf-8'))

• To request the license, please provide us with the machine code obtained from the getMachineCode function.

Please contact us:

🧙Email: [email protected]
🧙Telegram: @kbyaisupport
🧙WhatsApp: +19092802609
🧙Discord: KBY-AI
🧙Teams: KBY-AI

How to run

1. System Requirements

• CPU: 2 cores or more (Recommended: 2 cores)
• RAM: 4 GB or more (Recommended: 8 GB)
• HDD: 4 GB or more (Recommended: 8 GB)
• OS: Ubuntu 20.04 or later
• Dependency: OpenVINO™ Runtime (Version: 2022.3)

2. Setup and Test

• Clone the project:

  git clone https://github.com/kby-ai/FaceRecognition-Docker.git

• Download the model from Google Drive: click here

  cd FaceRecognition-Docker
  
  wget --load-cookies /tmp/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id=19vA7ZOlo19BcW8v4iCoCGahUEbgKCo48' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\1\n/p')&id=19vA7ZOlo19BcW8v4iCoCGahUEbgKCo48" -O data.zip && rm -rf /tmp/cookies.txt
  
  unzip data.zip

• Build the Docker image:

  sudo docker build --pull --rm -f Dockerfile -t kby-ai-face:latest .

• Run the Docker container:

  sudo docker run -v ./license.txt:/home/openvino/kby-ai-face/license.txt -p 8081:8080 kby-ai-face

• Send us the machine code and then we will give you a license key.

  After that, update the license.txt file by overwriting the license key that you received. Then, run the Docker container again.

  

  

• To test the API, you can use Postman. Here are the endpoints for testing:

  Test with an image file: Send a POST request to http://{xx.xx.xx.xx}:8081/compare_face.

  Test with a base64-encoded image: Send a POST request to http://{xx.xx.xx.xx}:8081/compare_face_base64.

  You can download the Postman collection to easily access and use these endpoints. click here

3. Execute the Gradio demo

• Setup Gradio Ensure that you have the necessary dependencies installed.

  Gradio requires Python 3.6 or above.

  You can install Gradio using pip by running the following command:

  pip install gradio

• Run the demo Run it using the following command:

  cd gradio
  python demo.py

• You can test within the following URL:
  http://127.0.0.1:9000

About SDK

1. Initializing the SDK

• Step One

  First, obtain the machine code for activation and request a license based on the machine code.

  machineCode = getMachineCode()
  print("machineCode: ", machineCode.decode('utf-8'))

• Step Two

  Next, activate the SDK using the received license.

  setActivation(license.encode('utf-8'))

  If activation is successful, the return value will be SDK_SUCCESS. Otherwise, an error value will be returned.

• Step Three

  After activation, call the initialization function of the SDK.

  initSDK("data".encode('utf-8'))

  The first parameter is the path to the model.

  If initialization is successful, the return value will be SDK_SUCCESS. Otherwise, an error value will be returned.

2. Enum and Structure

• SDK_ERROR

  This enumeration represents the return value of the initSDK and setActivation functions.

  Feature	Value	Name
  Successful activation or initialization	0	SDK_SUCCESS
  License key error	-1	SDK_LICENSE_KEY_ERROR
  AppID error (Not used in Server SDK)	-2	SDK_LICENSE_APPID_ERROR
  License expiration	-3	SDK_LICENSE_EXPIRED
  Not activated	-4	SDK_NO_ACTIVATED
  Failed to initialize SDK	-5	SDK_INIT_ERROR

• FaceBox

  This structure represents the output of the face detection function.

  Feature	Type	Name
  Face rectangle	int	x1, y1, x2, y2
  Face angles (-45 ~ 45)	float	yaw, roll, pitch
  Face quality (0 ~ 1)	float	face_quality
  Face luminance (0 ~ 255)	float	face_luminance
  Eye distance (pixels)	float	eye_dist
  Eye closure (0 ~ 1)	float	left_eye_closed, right_eye_closed
  Face occlusion (0 ~ 1)	float	face_occlusion
  Mouth opening (0 ~ 1)	float	mouth_opened
  68 points facial landmark	float [68 * 2]	landmarks_68
  Face templates	unsigned char [2048]	templates

  68 points facial landmark

  

3. APIs

• Face Detection

  The Face SDK provides a single API for detecting faces, determining face orientation (yaw, roll, pitch), assessing face quality, detecting facial occlusion, eye closure, mouth opening, and identifying facial landmarks.

  The function can be used as follows:

  faceBoxes = (FaceBox * maxFaceCount)()
  faceCount = faceDetection(image_np, image_np.shape[1], image_np.shape[0], faceBoxes, maxFaceCount)

  This function requires 5 parameters.

  • The first parameter: the byte array of the RGB image buffer.
  • The second parameter: the width of the image.
  • The third parameter: the height of the image.
  • The fourth parameter: the FaceBox array allocated with maxFaceCount for storing the detected faces.
  • The fifth parameter: the count allocated for the maximum FaceBox objects.

  The function returns the count of the detected face.

• Create Template

  The SDK provides a function that enables the generation of templates from RGB data. These templates can be used for face verification between two faces.

  The function can be used as follows:

  templateExtraction(image_np1, image_np1.shape[1], image_np1.shape[0], faceBoxes1[0])

  This function requires 4 parameters.

  • The first parameter: the byte array of the RGB image buffer.
  • The second parameter: the width of the image.
  • The third parameter: the height of the image.
  • The fourth parameter: the FaceBox object obtained from the faceDetection function.

  If the template extraction is successful, the function will return 0. Otherwise, it will return -1.

• Calculation similiarity

  The similarityCalculation function takes a byte array of two templates as a parameter.

  similarity = similarityCalculation(faceBoxes1[0].templates, faceBoxes2[0].templates)

  It returns the similarity value between the two templates, which can be used to determine the level of likeness between the two individuals.

4. Thresholds

The default thresholds are as the following below:

FaceRecognition-Docker/app.py

Lines 18 to 20 in 7580059

	verifyThreshold = 0.7