Member Info for TheLongestShot

Another layer on top of DMS.

It is like a virtual tour guide.
Given the planned route and the user's profile it will direct you to look at particular objects and know if you have spotted them, if not it can direct your gaze then it will recognise that you are tracking the object with your gaze as the vehicle moves
Interesting find JC

Re: BOSCH GMBH ROBERT [DE] - 2022-04-28Sun 10:28
SYSTEMS AND METHODS FOR DETECTING SYMPTOMS OF OCCUPANT ILLNESS

DavidW, this is not infringing on SEE's core territory.

This is actually a pretty neat or scary/Orewelian idea. Have a look at the photos in the patent and read the claims. This isn't really aimed at cars, but public transport where cameras/radar watch and/or listen to the passengers to see the signs of illness. Then the display (for driver/guard/security to be shown the potentially ill patient with the actions highlighted on the video - cough, sneeze, wretching, writhing. Only a small step from recording so called anti-social activity which is what worries me.

There is no reason for SEE to be linked with this

Seems to provide no function than see if the driver stops being sleepy or clumsy if you tell the not to.

No worry for us

Numpti,

Eyeware only do PC software and it is rather hard to hide the 3d gaze trackers with their limited range and refresh rate, that they support in a ****pit environment. Only really acceptable in mock-ups with large monitors not representative ****pits

It doesn't take long to see past their demonstrations and spot the gaps. They do automotive too - now where will they install the PC?

"buy indicator" has spoken

Phil, are Toyota developing their own DMS? Well, the diagrams of the Toyota patents may all be similar, but the group of names on this patent are the stalwarts of TRI that we know we have worked with for a long time . But as S2030 says, it is "developing" a DMS they are "probably" developing on top of our DMS.

This patent for example is about how to deploy the DMS - it turns out (sorry for the pun) but steering wheels move and block the camera. This novel implementation uses multiple fibre optics to take the light from individual lenses on the rim of the wheel facing the driver to a individual camera in the steering wheel of steering column. The theory being that at least one of the cameras that serves the top of the steering wheel will have a clear view of the face.

Of course this is completely impractical and expensive, not to mention that he lenses will be dirty, smeared or scratched and the image down a fibre will not be as clear as from a direct camera to lens arrangement.

This is the category of patent that exists as a road block, they have no intention of developing a solution to use it, but if every road is blocked this may be the least bad solution and they can reap the license fees from the fool that succeeds in making it work

Re: INTERFACE SHARPNESS DISTRACTION MITIGATION METHOD AND SYSTEM
JC, as you probably noticed this is from the same group at Toyota in the US and uses similar diagrams (car with a LIDAR and sensor roof rack)
It explains the basics of Eye tracking and dynamically blurs the infotainment screens if the driver and passenger aren't looking. When it sees them move the focus area it ensures the display is crisp when they look at it - the reason is to conserve valuable "brain juice" - it doesn't explain that the effect is probably lost by the world whizzing past full of advertising hoardings and other vehicles outside of the car!

"[0035] Provided the teachings in the paper by Hubel et al., the inventors have concluded that when sharp lines are present, the brain is engaged more, as opposed to when softer edges/lines are present. In other words, the brain uses more metabolic resources (i.e., blood, oxygenation, etc.) when processing sharp lines and edges versus softer lines and edges because detecting sharp lines and edges is a specialized function of a subset of neurons. Hence, when sharp lines are present in the peripheral view for task-irrelevant information, metabolic and thus, cognitive resources are being consumed unnecessarily. Such a circumstance is an inefficient use of metabolic resources, which should instead be allocated to task-supporting neural regions. An aspect is to soften image regions with edges and lines outside and away from the focus vision area. Blurred/softened lines do not lead to firing of the subset of neurons in the human visual cortex. An aspect is to gradually change from sharp image regions/lines to softened image regions/lines as the person shifts focus away from the focus vision area in an image. An aspect is to quickly sharpen image regions/lines in an image in anticipation of a person shifting their head to a new focus vision area. By softening image regions/lines outside of the focus vision area of a driver's vision, the cognitive resources freed up by blurring or softening image regions/lines of task-irrelevant images in the periphery can be reallocated to other brain regions necessary for supporting the driver, to use cognitive resources efficiently while driving a vehicle, especially when performing mission-critical tasks. An abrupt change from softer to sharper image regions/lines may help to draw attention in order to reduce time in focusing on a display image so that the driver can revert back to focusing on the task at hand. Conversely, a gradual, inconspicuous change from sharp to softer image regions/lines may minimize—or potentially eliminate—the drawing of attention to task-irrelevant information in the periphery so that the driver can continue focusing on the task at hand. Although the term “soften”, e.g., softened, softening, softens, softer, is used in this disclosure, it should be understood that the term “smooth”, e.g., smoothed, smoothing, smooths, smoother, could be used as well ..."

Re: COGNITIVE TUNNELING MITIGATION DEVICE FOR DRIVING
Hmm, Sometimes Toyota impress me, other times they disappoint, Here is the second claim in the patent:

2. The driver monitor system of claim 1, further including a first machine learning device,
wherein the audio-video device outputs a verification request of whether or not the driver feels drowsy or is focused on a point other than a driving task, and receives a response to the verification request, and
wherein the eye gaze direction, the eye lid position, the heart rate variability, and the response are fed back to the machine learning device which learns to predict whether the driver is transitioning into a cognitive tunneling state or a fatigue state.

So it asks you if you are tired or distracted, then learns to detect that state - I hope the driver tells the truth and doesn't just swear at the car!

JC, I rest my case.

What Lewis said!

That patent was tied to a particular hardware implementation that would be available for OEMs. But the implementation requires external processing. Which may not know the physical positions of the cameras, so it would not know if you were watching directly instead of via the mirror. Of course the lack of Camera in the mirror means there would be no DMS in this patent.

Could they add them, yes. But without full visibility of the drivers eyes nagging them to look at the mirror because they are reversing towards a parked car that they are looking at isn't helpful. When would a mirror mount DMS not see your eyes, when your body and head is twisted so that you can look out the rear window!

But I have no doubt

Multiple external cameras, but no internal ones or DMS/OMS

No, that sounds like their own implementation. Of course, it may only get installed in their own cars and even then maybe not in Europe for long

Both registered on the same day but although on similar themes, they are quite separate.

Of course we know that GM have Qualcomm for their future driving control so it is unlikely that the teams would be working together now, even if they were in 2020

Thanks JC,
Interesting patent, shows the thinking of NVidia. There is some black box magic that determines an attentativeness score based on saccades and gaze patterns; and a cognative load factor which compares past history.

But the crux is it will either warn you, or suppress an alarm based on whether you gazed at something outside.

So this is no direct worry for us, as the warn or suppress alarm part is not SEE, but part of the Tier1 implementation. Certainly we saw Veoneer demo this. Let's leave their lawyers to sort it out.

David, they key thing that you are missing is that the only Financial journalist with any nouse about DMS to ask any searching questions is Mr Safestocks

JC, I hope that if granted, Continental do what Volvo did with the 3 point seatbelt - don't charge for the patent to ensure it is used far and wide. Once a driver is detected to be inebriated or incapacitated, then the surrounding drivers and pedestrians deserve to have maximum protection from all of the safety tech in the car, whether the driver disabled it or not.

JC, this on one hand is a practical swap out of current lighting hidden above and behind the overhead luggage lockers in aircraft which mixes coloured RGB and IR LEDs. However, the usual shopping list of features is a hypothetical land grab. Yes IR cabin cameras exist and are useful. But claiming to perform eye tracking is a joke. This is by design indirect and diffused light, so could not provide a glint, just background lighting

Sorry JC, no need for SEE here, it is all about guessing what the person would want to look at and focusing or zooming as appropriate

DMS can help when the "Self drivers" are human and tired or distracted. But all of Seeing Machines valuable data, won't help "robots" to learn to drive cars. But it will help the humans who have to either monitor the driving all of the time - Tesla and level 2 to stay alert and aware of the surroundings

When Level 3 cars are allowed to drive within their ODD, DMS will ensure that the driver is ready to take over.

And for Waymo (and one or two others), Seeing Machines' bDMS (backup driver monitoring system) is currently ensuring that the safety drivers are staying alert while monitoring the test vehicles out on the road, so we don't have more Elaine Herzberg issues during testing.

Brick, perhaps you need to check myth number 3:
https://twitter.com/seeingmachines/status/1508308237405376514?s=20&t=R0iKpOCvYw7fVhJa5k-4xg

A handful of seconds of a video of a driver being distracted or having a microsleep won't solve "self-driving" problems.

It took 8,873,644,506 kilometres of monitoring drivers for SEE to record 11,592,646 distraction events.
In the 12 months they recorded 254,504 fatigue events that resulted in an intervention