# Face Recognition

## Introduction

Face recognition help you to matching a human face from an image. 

## Ability

* Recognize face(s) in all orientations
* Automatically detect and crop face(s) in the image

## Limitations

The following is a list of conditions that can reduce performance of the model to recognize face.

* Blurry image
* Noisy image
* Poorly lightning condition or light reflection (often flashlight)
* Partially hidden or obstructed faces

## Under the Hood

Kredibel is using **deep learning** to perform face recognition task. Our model achieve near **99.7%** accuracy in the labeled face in the wild (LFW) dataset. Below is an explanation of how our Face Detection system works.

### 1. Upload Image

Users can upload a **handheld** or **non-handheld** identity card image. Our system can handle both. 

### 2. Orientation estimation

Our system will estimate image orientation and normalize it. That's why our system can recognize faces in all orientations.

### 3. Face Detection

Our system will **detect** all face(s) in the image.

### 4. Cropping

Our system will automatically **crop** the face area detected by face detection model.

### 5. Face Recognition

Our system will **recognize** face(s) by extracting face unique features.

### 6. Face Comparison

We **compare** face unique features of the given image.

## Endpoints

| Method | URL                            |
| ------ | ------------------------------ |
| POST   | `$vision/v1/face/verification` |