Google Research has quietly made a breakthrough in the emerging area of image recognition and rapid video analysis — a breakthrough that has significant implications for pedestrian detection. Pedestrian detection refers to the analysis of statistics of pedestrian scale, occlusion and location. It gives detection systems a base of knowledge under which to operate.
Now on to Google's breakthrough.
As an experiment [PDF], Joe Yue-Hei Ng, Matthew Hausknecht, Sudheendra Vijayanarasimhan, Oriol Vinyals, Rajat Monga and George Toderici — researchers from the University of Maryland at College Park, the University of Texas at Austin and Google — applied several deep neural network architectures to analyze a dataset of videos over longer time periods than previously attempted.
It is a departure from the usual method of image-recognition analysis, which is convolutional neural networks, or CNNs.
The results were impressive, not just to the researchers ("our best networks exhibit significant performance improvements over previously published results…") but also to observers of this and related developments.
"At 15 frames per second, the Google team set a dramatic record that does not sacrifice either speed or accuracy against benchmarks like the Caltech Pedestrian detection metric," wrote Nicole Hemsoth at The Platform (hat tip to Hemsoth for first highlighting the study).
Pedestrian detection and prediction
The goals of the research are admirable and, OK, have the potential to be highly profitable. On that subject, Hemsoth writes that "use cases for how GPUs will power real-time services off the web are still developing. Pedestrian detection is one of those areas where, when powered by truly accurate and real-time capabilities, could mean an entirely new wave of potential services around surveillance, traffic systems, driverless cars, and beyond."
But as commenters to published accounts of this research project and other technologies deployed for pedestrian detection have noted, these emerging capabilities have scary implications as well.
First, though, a use case we can all get behind.
Driverless cars and pedestrians
It is widely acknowledged that advances in pedestrian detection will make driverless cars even safer, and also those driven by humans. However, there are limits to what the technology can do.
The California Institute of Technology, which established among the first and widely recognized datasets of such statistics, wrote in a paper [PDF] on the subject that "in the US, nearly 5,000 of the 35,000 annual traffic crash fatalities involve pedestrians — hence the considerable interest in building automated vision systems for detecting pedestrians."
It tested 16 "pre-trained state-of-the-art detectors across six datasets" on the market, and to make a long story short, found that "even though progress has been made, there is still much room for improvement."
"In particular, detection is disappointing at low resolutions and for partially occluded pedestrians," the Caltech paper concluded.
Google's research in this area suggests that it could address these and other issues with the current modes of measuring pedestrian traffic, at least when videos are used.
The privacy factor
But as this happens, privacy is going to be an issue, as other technologies currently being applied to this issue suggest.
John F. Kennedy International Airport in New York is trying out a new system that uses passengers' mobile phones to get a sense of how long people are taking to go through the security and other lines.
It uses technology developed by Denmark's Blip Systems.
According to Giz Mag, which picked up the story first:
The technology utilizes beacon modules that detect the Wi-fi or Bluetooth signal of passing mobile devices which are in "discoverable" mode. When a device is discovered, the system records, encrypts and time-stamps its MAC address — this is an ID that’s unique to that mobile device. None of the user’s personal data is accessed or recorded.
When that same phone or tablet is re-identified by other beacons farther down the same queue, the system analyzes how long it took to get from one beacon to the other. It then extrapolates how fast it will take someone to get through the line, and displays that information on a nearby screen.
"Yeah right," was, essentially, commenters' response to this story and Gizmodo's report of the technology as well.
A sampling:
LOL "None of the user’s personal data is accessed or recorded." - this is an airport - if they were not recording this, they would be criminally negligent at their job. Imagine the uproar if an attack happened, but they'd used BLE MAC's for wait queues, and not bothered to use SS7 etc to ID you, grab your texts, data, and calls - which all essentially use the exact same bit of $200 hardware to do?
I’m thinking that the TSA can now produce a nice list that correlates your passport or other identification given at the security gate with your Mac address. What they do with that list after the fact is anyone’s guess. It just means they don’t have to go after Verizon, AT&T, T-Mo and Sprint for that info.
"Tireless Eyes"
Even Hemsoth's excellent review of what Google is doing raised eyebrows among the readers. NextBigFuture reposted the article.
Said one commenter:
Going to need something like this, to process all the data coming in from new operating systems turning all our connected hardware into Orwellian "telescreens." Lack of enough eyes, especially loyal, tireless ones, has always been the chief obstacle to panopticon surveillance.
Okay commenters, point made: Advances in pedestrian detection have serious implications for privacy and government surveillance. And as for me personally, I really don't care to be displayed a personalized ad while I am driving by a digital sign, which is possible given Google's involvement. It could distract me and cause an accident.