Augmented reality system filters out moving objects

The folks who created the amazing security camera screen saver (now available for Windows), are working on a real-life version of the "superplonk" reality filtering system that author Charles Stross introduced in his terrific novel, Accelerando.

 Wp-Content Uploads 2008 01 Superplonk

In Accelerando Charles Stross writes about a lot of interesting concepts we are just starting to work on. It was one of the most important books for me in 2007. it shows how close science fiction and science get in these days.

My favorite feature is "superplonk." It remixes the environment and filters annoying persons, objects and sounds. That’s an augmented reality version of what I practice today with special earplugs. But soon that should be possible with modified hearing devices and slim head mounted displays.

One experiment in my ongoing surveillance series simulates superplonk with images of network cameras. Via motion detection I am reconstructing a place’s image without people and cars. All moving objects are becoming ghosts. Only people and cars who are standing still are becoming visible. Movement makes you invisible. Jan covers this topic in his master thesis, too.



  1. Motion equaling invisibility was an early issue with photography, due to long exposure times. This photo taken by Louis Daguerre in Paris in 1838 or 1839 is the first image of human being permanently captured. He’s in the lower left, with his leg raised. The photo should show a busy street, but the motion of the people and carriages have made them vanish into time. The image of the one man survives only because he is standing still, probably having his boots blacked by a phantom street urchin.

  2. You can already do this in Photoshop CS3 by converting the layers of a psd containing all the frames to a smart object and changing the stack mode to median.

    Here’s an example:

  3. Somebody wrote a master’s thesis on this? Wow.

    I was messing with PySight on my Mac a few months ago and wrote a filter that does this with video from the iSight camera. It looks for pixels that change and figures out what the background is. So if you fire it up while sitting in front of your computer you appear until you start to move. Then parts of you disappear as you reveal the background. New people walking into the room never appear unless they hold still for a while and are detected as background.

  4. Does anyone else think this could have a real-world application as a way to weed out terrorists and other ne’er-do-wellers loitering around buildings to scope out their security systems and accessways?

    Or have I just seen way too many spy movies?

  5. “Nobody can be so amusingly arrogant as a young man who has just discovered an old idea and thinks it is his own.”

    Didn’t BB report a while back on a site where you can upload a number of photos taken of the same view, and it remove everything in motion?

  6. …Actually, I use this trick all the time when taking pictures of outdoor attractions, where inconsiderate idiots and/or their festering excuses for offspring are oblivious to and/or clearly aware of the fact that I’m trying to take a picture and walk in front of the camera. Just take the 2-5 images you need, slap them in Photoshop as layers, and then erase where necessary. Eventually you’ll have a photo of said attraction without any dipshits cluttering the beauty of what you were shooting.

    Works best with a tripod, but a steady hand can do the job.

  7. Not to spoil the fun, but I believe this may have also been included in the program EffecTV (or was it FreeJ?)

    Still, this discussion provided plenty of information that I was not aware of previously, like that fancy Photoshop trick…

  8. Do the BBers actually have to invent what they post? And invent it right before they post it? Oh, oh, I take that back: invent it right before you read it? Is that the new standard, because I missed that memo.

  9. I went to a Mitre corporation open house almost 25 years ago and saw a live demo of this. you had to stand still for like 20-30 seconds before you showed up on the monitor. and moving your arms about slowly made the image really cool. Not sure how you can get a master’s thesis out of ancient technology, but I’m sure there’s always a new feature.

    Antinous, you might need to relax before you notice that everyone on the Internet is wrong…

  10. Antinous, you might need to relax…

    Someday I’m going to find that dictionary….

    1. not disagree with

  11. hi there,
    we used this “technique” for some of our works. Only non and few moving objects and people manifest in the pictures/videos. We don`t want to filter every person out but try to find something inbetween by adjusting the variables in our programming (max/msp/jitter). We also refer to to Daguerre photo mentioned above. Take a look if you like:





Comments are closed.