Simple Face Recognition

A face recognition network can be trained in such a way that the Euclidean distance in the embedding feature space directly corresponds to face similarity. Using the embedding as facial descriptors, we can implement a simple face recognition algorithm without the need to train a new model.

Here is a small sample of a set of images from a family of five.

In[1]:=1

Compute the facial descriptors for each person.

Copy to clipboard.

In[2]:=2

✖

Direct link to example

You can use FeatureSpacePlot to visualize the descriptors in clusters.

Copy to clipboard.

In[3]:=3

✖

Direct link to example

Out[3]=3

Use Classify to perform classification in the feature space.

Copy to clipboard.

In[4]:=4

✖

Direct link to example

Apply the classifier on a new image.

Copy to clipboard.

In[5]:=5

✖

Direct link to example

Copy to clipboard.

In[6]:=6

✖

Direct link to example

Out[6]=6

Visualize the classification on top of the test image.

Copy to clipboard.

In[7]:=7

✖

Direct link to example

Out[7]=7

Try the classifier on an image taken at a different point in time.

Copy to clipboard.

In[8]:=8

✖

Direct link to example

Copy to clipboard.

In[9]:=9

✖

Direct link to example

Out[9]=9