Tracking and bluring faces in multiple 360 images via python opencv

Question

Is there a way to track down and nicely blur faces or part of face (like hair) for multiple 360 degree images via python opencv. ? I'am using Windows OS and python3.8

Answer 1

Yes, there is. First you need to detect the face(s) using Haar-cascade , which will provide you the rectangle coordinates of the face location. Then you can use this answer to blur the desired portion of an image.

Answer 2

Two methods with opencv and python

1. Using a Gaussian blur to anonymize faces in images and video streams
2. Applying a “pixelated blur” effect to anonymize faces in images and video

The method is well explained here and you can access code.

Now, a more advance solution if you are using GPU, and you want to run the application on a live video stream.. its with nvidia DS and Deep Learning. The github here reports results on T4, i believe you should be able to run it on Jetson nano. Here is the link