Face-Recognition-Deep-Learning

Face recognition is used for to unlocking cell phones. And with recent advancements in deep learning,In this repository how to develop a face recognition system that can detect faces in images, identify the faces, and even modify faces with "digital makeup" like you've experienced in popular mobile apps

Part 01 - Face Recognition Pipeline Steps

Step 1: Locate and extract faces from each image Step 2: Identify facial features in each image Step 3: Align faces to match pose template Step 4: Encode faces using a trained neural network Step 5: Check Euclidean distance between face encodings

Use of face recognition

1. Identify verification’
2. Automatically organizing new photo libraries by person.
3. Tracking a specific person
4. Counting unique people
5. Finding people with similar appearances.

Install libaries

pip install pillow 
pip install face-recognition

if you are using google Co-lab set to Run time to GUP

Part 02 - Face Recognition

Sliding window classifier

Histogram of Oriented Gradients (HOG)

Analyzing an Image as a Histogram of Oriented Gradients

face land mark estimation

Identify key pint on the face – tip of the nose , center of the eye

Identifying Face Landmarks with a Machine learning model

Calculate affine Transform

Finally -

Part 03 - Representing a face as a set of Measurements

measure each eye, size of the cheekbones and the width of mouth and so on

Here are two face , lets take three measurements for each face i. Length of nose – 2.5 (left) | 2 (right) ii. Width of mouth – 2.5 and 3 iii. Distance of one eye to other 4 and 3.5

That distance is relay small or close each other(two point) that may be a same person.

using face_encodings library - output 128 values

Part 05 Euclidean distance

Face distance threshold – Set a face maximum distance that is still considered the same face . Lets assume the threshold value is 0.6.

• If the distance (a,b) > 0.6, not , match
• If the distance (a,b) =< 0.6, they match
• he lower the distance , the better the match
  
  

Part 6 - Fun Uses of face recognition

• How to drawaing a image (makeup)
• Choose best image from collection of images