How to blur a new face with Secure Redact
Many people have tonnes of video they need to process for different reasons, e.g. video analytics, DSARs, security etc. Secure Redact is a quick and easy tool to redact personal data from videos, including faces, tattoos, license plates, and vehicles so that video data can be used and shared without compromising privacy.
Secure Redact automatically detects the faces, heads and license plates of the video, ready for redaction. Once you have uploaded your media, this will be shown by the orange boxes. These are the automatically added redaction boxes and where, in the exported video, the faces and license plates will be blurred.
From here, you have lots of different options for the review process. In this case, we are going to talk through a common usage: how to blur a new face.
There may be the case that the system hasn’t added an orange redaction box to a particular face that you need to blur. This is because the Secure Redact system cannot be 100% accurate. Videos are long, have lots of movement, and can have a lower resolution quality or occlusion from different camera positioning. As a result, automatically redacting ALL the faces is tough - even for AI.
For example, in an hour of video from a busy city street, you could have roughly 2 million faces in the video. Even a system that was 99.9% accurate would still miss 2,000 faces.
Not to worry, we have a quick and easy solution.
First, you need to locate where the individual you want to blur appears in the video. You can go through the video by using the video controls at the bottom of the screen. Once you’ve found the right frame where the individual appears, you are ready to go.
Right click anywhere within the video player and then click add track.
Quick fire: what is a track?
A track is a consecutive set of redaction boxes that follows the same face (or number plate) throughout the video. It allows you to have consistent redaction of that one individual across multiple frames.
You should then see a yellow box that you can adjust the size of, and drag to cover the face you need to redact.
You will see the individual track view appear on the right, with the start and end frame numbers.
These will be the numbers of the first and the last frame in which your individual appears. You’ll notice that the start frame is now the same as the frame number you are currently on (i.e. where you added the yellow box).
You can then scroll through the frames by going forward in the video with the video controls - up until where you want that track to end (i.e. when the individual goes out of frame). Then right click on the screen again, and click “Add track end”. Alternatively, once you have found the end frame, press the “+” button next to the end frame box in the individual track view.
This will make another yellow box appear. Again, adjust the box as required and place it on your individual’s face.
Quick fire: what is the difference between a track and a zone?
While a track is a consecutive set of redaction boxes that follows the individual throughout the video, a zone stays within a specific area of the video screen. For example, you could draw your yellow track box over a corner of the screen, then choose it as a zone, and it would stay in that corner and blur whatever came into that box, throughout the video.
Once you have chosen “track” or “zone”, the “start” button will illuminate. Press “start” and that track will move into the processing tab.
Once a track has been processed, its yellow box will turn orange and be added to your tracks panel.
You can double check the track follows your individual correctly throughout all the frames that they appear by finding the specific track in the tracks panel, clicking the arrow next to the start frame, and then press play.