YouTube's improved Blur Faces tool launches
Ryan Stevens, Software Engineer at YouTube, announced today that YouTube has updated the service's Blur Faces tool to improve its accuracy and ease of use.
YouTube introduced a face blurring tool back in 2012 as a feature to anonymize faces in videos. The feature blurred all faces that the algorithm detected automatically which limited the use of the feature significantly.
The Blur Faces tool was improved in February 2016. The new version of the tool supported the blurring of any object in videos, even if they moved around.
The August 2017 update to the Blur Faces tool improves it even further. When you open the tool now on YouTube, you get a list of faces that Google's algorithm detected in the video.
All you need to do at this point is to select the faces that you want to appear blurred in the video. YouTube's algorithm is capable of recognizing the same person across the video according to Google.
The tool is designed for a wide array of situations that we see in YouTube videos, including users wearing glasses, occlusion (the face being blocked, for example, by a hand), and people leaving the video and coming back later.
YouTube editors can access the feature in the following way:
- Open the Creator Studio on YouTube, and there the video manager.
- Select to edit the video that you want to blur faces in.
- Switch to the Enhancements tab.
- Locate and select the "blurring effects" sub-tab on the page that opens.
- Select the edit button next to blur faces.
Note that YouTube may need to process the video for the feature. This processing may take a while to complete. Google's algorithm scans the video for faces during the processing basically.
During processing, we break your video up into chunks of frames, and start detecting faces on each frame individually. We use a high-quality face detection model to increase our accuracy, and at the same time, we look for scene changes and compute motion vectors throughout the video which we will use later.#
Once we’ve detected the faces in each frame of your video, we start matching face detections within a single scene of the video, relying on both the visual characteristics of the face as well as the face’s motion.
You may exit the page and do something else on YouTube until the processing is done.
YouTube lists the faces that the scan detected afterwards. You may then blur one, multiple or all faces in the video. The video is embedded on the same page so that you can play it to test the effectiveness of the procedure right away.
Hit the save button once you are done. Ryan Stevens notes that videos won't lose views, likes or comments when you apply blur to faces and use the save option.
Note that you need to use "save as new" and delete the original video if you want to remove the original video from YouTube.
Stevens suggests that editors preview the video before they publish it with the blurred faces, and that editors use the custom blurring feature to correct any issues that may come up.Advertisement