Lambda Labs Face Recognition API Introduction

We hate long and boring documentation too. Let's skip right to the code examples (the full documentation starts below the horizontal line). So here's facial recognition in just Four Easy Steps.

Step 0. Sign up for the API

Step 1. Create an Album

An album is a set of tagged images that can be used to recognize new ones. Let's instantiate an api and create an album. (Album names are globally unique, so you'll have to choose something other than 'Presidents'.)

from FaceRecognition import FaceRecognition 
api = FaceRecognition("YOUR_PUBLIC_KEY", "YOUR_PRIVATE_KEY")
# create an album 'Presidents' and save album_key
album = "Presidents"
album_key = api.createAlbum(album).body['albumkey']

Step 2. Train the Album

Give the album a few labeled images to work with. The more images you give it, the more accurate the model will be. Note, you can give us either urls or files.

# train the 'bush' label (two urls and one file)
api.trainAlbum(album, album_key, 'bush', None, "http://lambdal.com/bush/0.jpg") # using variable arguments
api.trainAlbum(album, album_key, 'bush', urls="http://lambdal.com/bush/1.jpg") # using kwargs
api.trainAlbum(album, album_key, 'bush', files=open("bush2.jpg"))
...

# train the 'clinton' label (using all files [but two ways to call it])
api.trainAlbum(album, album_key, 'clinton', open("clinton1.jpg"))
api.trainAlbum(album, album_key, 'clinton', open("clinton2.jpg"))
api.trainAlbum(album, album_key, 'clinton', files=open("clineton3.jpg"))
...

Step 3. Rebuild the Album

Tell the album to build a model itself using the labeled images.

# rebuild the album after training
api.rebuildAlbum(album, album_key)

Step 4. Recognize an Image

Give the album an image to recognize. You can see that there are two options, urls and files.

# run recognition using a url
link_response = api.recognize(album, album_key, urls="http://lambdal.com/clinton/19.jpg")  
# run recognition using a file upload
file_response = api.recognize(album, album_key, files=open("clinton.jpg"))
print("Bill Clinton was classified as:")
print(link_response.body['photos'][0]['tags'][0]['uids'])

Sign Up
Get Keys
Download Client
Code Examples
Current & Upcoming Functionality
Support

1. Sign Up

Select a plan to access our API.
Create an account

Detect

from FaceRecognition import FaceRecognition 
api = FaceRecognition("MY_PUBLIC_KEY", "MY_PRIVATE_KEY")
# File Upload
api.detect(files=open("~/Desktop/face.jpg"))
# Url Link
api.detect(urls="http://lambdal.com/test2.jpg")

require File.join(File.dirname(__FILE__), "FaceRecognition.rb")
api = FaceRecognition.new("MY_PUBLIC_KEY", "MY_PRIVATE_KEY")
# File Upload:
api.detect(File.open("/Users/sabalaba/Downloads/jennifer.jpg"))
# URL link:
api.detect(nil, "http://lambdal.com/tiger.jpg")

require_once("FaceRecognition.php");
$obj = new FaceRecognition("MY_PUBLIC_KEY", "MY_PRIVATE_KEY");
// File Upload
$obj->detect(fopen("~/Desktop/face.jpg", "r"), NULL);
// URL Link
$obj->detect(NULL, "http://lambdal.com/test2.jpg");

import FaceRecognition;
import java.io.File;
// and when you are ready to use it just create a new instance!
FaceRecognition client = new FaceRecognition("MY_PUBLIC_KEY", "MY_PRIVATE_KEY");
client.detect(new File("~/Desktop/face.jpg"), null);

Recognize

from FaceRecognition import FaceRecognition 
api = FaceRecognition("MY_PUBLIC_KEY", "MY_PRIVATE_KEY")

# create an album and save album_key
album = "Presidents"
album_key = api.createAlbum(album).body['albumkey']
print "Got Album Key: ", album_key

# train the bush label
api.trainAlbum(album, album_key, 'bush', urls="http://lambdal.com/bush/0.jpg")
api.trainAlbum(album, album_key, 'bush', urls="http://lambdal.com/bush/1.jpg")
api.trainAlbum(album, album_key, 'bush', urls="http://lambdal.com/bush/2.jpg")
...

# train the clinton label
api.trainAlbum(album, album_key, 'clinton', urls="http://lambdal.com/clinton/0.jpg")
api.trainAlbum(album, album_key, 'clinton', urls="http://lambdal.com/clinton/1.jpg")
api.trainAlbum(album, album_key, 'clinton', urls="http://lambdal.com/clinton/2.jpg")
...

# rebuild the album after training
api.rebuildAlbum(album, album_key)

# run recognition on a new image
link_response = api.recognize(album, album_key, urls="http://lambdal.com/clinton/19.jpg")  
file_response = api.recognize(album, album_key, files=open("clinton.jpg"))
print("Bill Clinton was classified as:")
print(link_response.body['photos'][0]['tags'][0]['uids'])

require File.join(File.dirname(__FILE__), "FaceRecognition.rb")
api = FaceRecognition.new("MY_PUBLIC_KEY", "MY_PRIVATE_KEY")

# Create an album and save the album key
album = "Presidents"
album_key = api.createAlbum(album).body['albumkey']
print "Got Album Key: ", album_key, "\n"

# train the bush label         
api.trainAlbum(album, album_key, 'bush', nil, "http://lambdal.com/bush/0.jpg")
api.trainAlbum(album, album_key, 'bush', nil, "http://lambdal.com/bush/1.jpg")
api.trainAlbum(album, album_key, 'bush', nil, "http://lambdal.com/bush/2.jpg")
...

# train the clinton label      
api.trainAlbum(album, album_key, 'clinton', nil, "http://lambdal.com/clinton/0.jpg")
api.trainAlbum(album, album_key, 'clinton', nil, "http://lambdal.com/clinton/1.jpg")
api.trainAlbum(album, album_key, 'clinton', nil, "http://lambdal.com/clinton/2.jpg")
...

# rebuild the album after training
api.rebuildAlbum(album, album_key)

# run recognition on a new image
link_response = api.recognize(album, album_key, nil, "http://lambdal.com/clinton/19.jpg")
file_response = api.recognize(album, album_key, File.open("clinton.jpg"))