skywatchai 0.0.1

An API Wrapper for powerful face detection, verification and recongition for python

Skywatch.ai - Facial Recognition System

Skywatch.ai is an API wrapper for powerful face detection and recognition models. It enables an efficient and clean way to use these models in your project without having to worry about the backend clusters. The focus of these project is to provide a easy-to-use package that speeds up development of lot of applications and research.

Usage

Face Detection

import skywatchai.SkywatchAI as skai
img_path = 'test/oscar.jpg'
detected = skai.detect_faces(img_path)

Face Verification

img1 = 'test/image1.jpg'
img2 = 'test/image2.jpg' 
result = skai.compare(img1, img2)
print('Are they same person:', result)

Face Recognition

Building the Database

For Face Recognition, you need to build the repository of people face images. You can have multiple images of same person under the directory of his name. Please refer to the below directory tree.

database
├── people
|   ├── Brad Pitt
|   |   ├── image1.jpg
|   |   ├── image2.jpg
|   |   ├── ..
|   ├── Bradley Cooper
|   |   ├── image1.jpg
|   |   ├── image2.jpg
|   |   ├── ..
|   ├── Chris Hemsworth
|   |   ├── image1.jpg
|   |   ├── image2.jpg
|   |   ├── ..
|   ├── ..

import skywatchai.SkywatchDB as skdb
# Give face_path directing to folder containing images following the above requirement
skdb.build_db(face_path='database/people/', save_path='database/')
faceDB, nameMap = skdb.load_db(path='database/')

Recognizing the person from Database

annot_img = skai.find_people(img, faceDB, nameMap)

Dependencies

• MTCNN
• FaceNet
• Tensorflow
• Keras
• Numpy
• OpenCV

Acknowledgement

I am very thankful for deepface library created by Sefin Seringil. His works were very useful for me in creating this project.