Face Detection

With Uploadcare Image Transformations, you can detect faces in your images for automatic cropping or blurring those out. Face Detection is implemented in line with our general Image Transformations workflow, as the detect_faces URL directive.

/:uuid/detect_faces/

Note, detect_faces is not divided from a file UUID by the /-/ separator. Hence, it can not be piped other operations.

detect_faces returns coordinates of faces found in an input image. The output is similar to the json operation. The output is a JSON with the additional list faces that holds coordinates of faces that were detected. Data for each of the found faces are put into separate lists that look like this:

[x, y, x_size, y_size]

Further, lists within faces contain:

• x, y — coordinates of the upper-left corner of an area where a face was found.
• x_size, y_size — dimensions of that area.

For example, let’s run a face-check for the following image on our CDN,

To do so, we put detect_faces into the image URL separating it with the forward slash / from the UUID:

https://ucarecdn.com/5128ec65-9957-47b8-a6ad-4c2f172ef660/detect_faces/

You will get the following faces list in the response JSON:

"faces": [
  [45, 142, 207, 207],
  [460, 113, 238, 238],
  [892, 43, 265, 265]
]

Technically, the operation detects faces using Haar Cascades. That approach deals with machine learning processes which rely on classifiers holding cascades of features specific to faces, eyes, etc.

detect_faces uses an algorithm that tends to better detect fronts of faces, rather than facial profiles. Also, covering important face features with different objects leads to a decline in detection accuracy.

The detect_faces operation is still subject to testing and recalibrating the backend when needed. So, if you ran into any issue with the operation or simply got questions, make a post in our community area.