Subscriber Discussion

Masking Software For CCTV Playback. Any Good Ones??

Is there a Masking software that you can recommend so that faces or cars can be masked out in playback.

Actually, I am not sure anything like that exists in production VMSes. It's certainly possible to do this but getting a commercial offering will not be easy. Even if you could get a third party offering, it's not clear how to integrate it smoothly.

How would the masking software know that the object in question that you want to anonymize is a face or a car? Anything that could actually do what you are asking would have to use some type of analytic to first determine what objects are.

Everything that I've ever seen for pixelating regions in video (to anonymize faces, or even cars if that's what you want) is a post-production action. You shoot the video first, then import the clip into an editing tool that allows you to add pixelated regions (as a layer) on top of the original clip.

Note: If the object in question is moving, this becomes a much more time-consuming effort as you have to review your whole clip and adjust the crop points of the blur pane. The faster the object moves, the more work you have to do.

Marty, detecting bodies and cars is fairly easy to do with analytics. So, yes, you would need analytics but it would be fairly straightforward. The problem is where to insert the analytic. Ideally, you would want it embedded in the VMS/recorder so it instantly mask and allow unmasking. Unfortunately, harder any of those systems supports this.

How fast can analytics analyze?

Wouldn't the object have to appear (at least for a small amount of time) in the field of view before analytics could detect that it is an object it should blur? And the busier the scene, the longer it should take for an object to be identified as an object that needs blurring, no?

If this is the case, then the 'playback' would have to crop off all footage that was recorded before analytics detected the objects to blur.

It's playback so it does not need to be real time. The analytic can scan the video in multiple passes to make sure the mask is properly and fully applied.

Btw, I doubt they would crop or destroy portions of the video. They simple can add an overlay mask on top of the video that tracks / sits above the objects. That way, if needed, they can turn off those masks, when needed, with the proper authority, etc.