Slashdot Mirror

← Back to Stories (view on slashdot.org)

UCF Researchers Perform World's First Automated Mass-Crowd Count (ucf.edu)

Posted by timothy on from the national-park-service-thrilled dept.

subh_arya writes: Automatic crowd counting has been an extremely challenging computer vision problem. However, researchers from UCF, seem to have found a reasonably accurate solution using sophisticated probabilistic models. Although there has been several previous efforts in this direction, this is the first time the technology has been put to use on a realistic scenario where around 550,000 protesters participated for Catalunyan Independence. A freely available technical paper published in IEEE Trans. on Pattern Analysis and Machine Intelligence, 2015 is available here.

21 of 32 comments (clear)

  1. This must be a hoax by damn_registrars · · Score: 4, Funny

    We learned years ago that people from Florida don't know how to count.

    --
    Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
    1. Re:This must be a hoax by Hartree · · Score: 1

      What? You mean UCF isn't University of California Fresno? :)

    2. Re:This must be a hoax by whopis · · Score: 3, Funny

      The event where people were protesting for Catalunyan independence was chosen specifically because there a very low likelihood of anyone name "Chad" being part of the protest.

      Floridians only have trouble counting when there are excessive Chad's hanging about.

  2. Yankee Nine-Niner... by freeze128 · · Score: 2

    "Food riot in progress. Approximately 1500 civilians. No weapons evident."

  3. Mooo? by PPH · · Score: 1

    Can't you just RFID tag their ears?

    --
    Have gnu, will travel.
    1. Re:Mooo? by Anonymous Coward · · Score: 1

      RFID-tagged ears are for COWS!

      Oh, wait...

  4. Food riot in progress,approximately 1500 civilians by xxxJonBoyxxx · · Score: 1

    First application would seem to be hooking this up to a system to automatically dispatch a drone to monitor if not disperse any detected crowd. Somewhere someone's salivating...

  5. Numbers by opkool · · Score: 1

    Writting "around 500,000" is quite controversial: the BBC article you link points to authorities' provided numbers.

    "Local police said 1.4 million people turned out but the Spanish government put the figure at no more than 550,000".

    So either you don't say anything or provided the number provided by the article.

    1. Re:Numbers by swb · · Score: 2

      When there's no provable number involved, everybody uses the number that suits their own interests. How else can you explain estimates that vary by 3x?

      Police? If the issue isn't police malfeasance, they like the higher number. If there was a riot, the large number supports their "need" for more police officers, weapons and expanded powers (greater leniency on the use of violence, more intelligence gathering, etc). If the protest was non-violent, they get to claim the credit for successfully managing it and keeping it peaceful.

      Government? They usually like the lower number, as it allows them to claim the organizers or issues lack political support so they can keep doing whatever it is they want.

      Protesters? They like the bigger number, it shows how much support they have.

    2. Re:Numbers by subh_arya · · Score: 1

      Writting "around 500,000" is quite controversial:

      I don't think I wrote "around 500,000". AFAIR, I quoted the number from BBC.

      --
      A computer scientist is someone who, when told to Go to Hell sees the "go to" rather than the destination, as harmful.
    3. Re:Numbers by moeinvt · · Score: 1

      "everybody uses the number that suits their own interests."

      Very true, and don't forget the media. They always exaggerate or underestimate numbers based on their particular biases.

      BLM protest++
      Gun rights rally--
      etc.

  6. Getting there... by argStyopa · · Score: 1

    Great!

    Now all we need to do is get that mass facial recognition tech enabled, and dovetailed with the personal-info databases in the NSA, then we don't have to "worry" about "protesters" ever again!

    --
    -Styopa
  7. Who needs drones WHEN THE SCOOPS ARE COMING! by swb · · Score: 1

    The scoops are coming!

    Your protest for more Soylent Yellow means you'll be turned into Soylent Green!

    Soylent Green is People!!!

  8. TMBG algorithm by sh00z · · Score: 1

    It's easy! Just count arms, legs and heads, and divide by five.

    1. Re:TMBG algorithm by Archangel+Michael · · Score: 1

      These are sheeple, so no arms. But your math still works fine. Excellent algorithm

      --
      Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
    2. Re:TMBG algorithm by Obfuscant · · Score: 1

      It's easy! Just count arms, legs and heads, and divide by five.

      News Item: "AP Next year's paralympics competition has been cancelled due to low numbers of participants and spectators. ... A quick survey of raised hands in the audience when asked 'would you attend next year' supports the declining numbers ..."

      On the actual topic, this is not a "count". It is a statistical measurement. You "count" the number of things your statistical operations identify as "person", but it's still statistics. That would make it an inappropriate way of conducting a census, for one thing.

  9. Human detection ? by swell · · Score: 1

    Opening line from UCF: "Human detection in dense crowds is an important problem..."

    I have never found difficulty detecting humans in a crowd. OK, nit picking, but I have a point.

    Skimming TFA I sense that much mathematical rigor went into this work, which is why I love science. But the opening sentence , above, is a dismal start in presenting the work. There is more convoluted verbiage following that, all of which suggests an eager young person hoping to impress and earn a Masters degree.

    Communication is essential. The best science, the best software, the best iDevice is useless unless people can understand it. DaVinci communicated via drawings remarkably well. Words can make understanding easy or painful, but that's what we are mostly required to use.

    I was a technical communicator. My job was to make concepts understandable to humans unfamiliar with them. I found it a pleasant challenge. But as I look around I see brilliant work being done in science and software and either no effort or ineffective effort to describe what it does/means. OSS is guilty, of course.

    Often the team member responsible for communicating is left out of the process. Sometimes they are considered to be a lower life form and shunned. But mostly that person does not exist. Consider that in your next project so that you don't look like this Masters candidate when you proudly present the results of your work.

    --
    ...omphaloskepsis often...
    1. Re:Human detection ? by Obfuscant · · Score: 1

      I have never found difficulty detecting humans in a crowd. OK, nit picking, but I have a point.

      He didn't write that it was a difficult problem for human beings, he wrote that it is an important problem in the context of computer vision. What you find difficult or easy can be exactly the opposite for computer vision, and whether a problem is important to solve or not has nothing to do with how easy it is for you to do it manually.

      For example, I have a current remote sensing problem that is important for us to solve. It is easy for a human being to look at the data and determine the answer. Unfortunately, I have neither the money nor the time to hire people to look at the hundreds of thousands of images to do the easy job, so I have to come up with a way of getting the computer to at least winnow that set down to a manageable number so humans can do the rest.

      Now, you may not consider this an important problem because you don't care, but it is important to the people who need to use the data. Just like it is important for the computer monitoring the camera overlooking the park to be able to signal to the human operators that there is a crowd of greater than the defined size gathering there and they might need to do something, even if that "something" is as simple as having the porta-potty people drop off a few more.

  10. Re: Food riot in progress,approximately 1500 civil by xxxJonBoyxxx · · Score: 1

    >> Why would a drone dispersing a crowd need an accurate head count? If the crowd is causing a problem or upsetting some one with power

    I think you answered your own question: if you can disperse a crowd BEFORE it causes a problem or upsets someone with power there's a business case to be made.

  11. Re:Maybe we can... by KGIII · · Score: 1

    I haven't read the paper and I'm still (and will be for a long while to come) rather wrapped up in an NDA but the end result (perhaps the process) is not new by any means. It may be more accurate and I'm sure the process is different but there has been machine automated crowd counting (post event) for quite some time. It wasn't easy or fast but it's been done before. In fact, some of the methods used came from privately funded research from a school not that far away.

    I'll definitely give the paper a read but it's going to be a while before I can sit down and do it any justice. I wonder what it cites? It's absolutely not entirely original. I don't see how it can but but I've neither read the abstract, paper, or the likes though I've probably read some of the work it's building from. I wish I'd a bit more time to devote to it while this thread is still active - I could give a more accurate opinion on the quality and originality.

    Consider, if you will, I've not been in the industry for eight years and it wasn't entirely new when I sold and retired. So... Yeah... Post-processing crowd count (estimate really) wasn't anything new. Live and real time might be new and interesting - also a huge amount of data. I'm assuming this is more accurate, uses a different methodology, and involves some serious data crunching.

    --
    "So long and thanks for all the fish."
  12. How many? by manu0601 · · Score: 1

    Before we had quotes such as "50000 persons according to protesters, and 30000 according to the police".

    Now we will have "50000 persons according to protesters, 30000 according to the police, and -1.045e18 according to the computer."