I've done a ton of mobile gaze tracking. We never went for the most important application here: babies will preferentially look at different things on a screen if they predisposed to autism. A screening tool is the easiest thing to make from a technical point of view and also the most useful for society. Why don't you try that? Current methods wait until the baby can talk and this could trigger intervention a very critical year earlier.
They create this CNN for exactly this task, autism diagnosis in children. I suppose this model would work for babies too.
Edit: ah I see your point, in the paper they diagnose autism with eye contact, but your point is a task closer to what my model does. It could definelty be augmented for such a task, we’d just need to improve the accuracy. The only issue I see is sourcing training data might be tricky, unless I partner with some institution researching this. If you know of anyone in this field I’d be happy to speak with them.
That's great! What I'm talking about is a bit different though and might be a lot easier to deploy and work on much younger subjects:
Put a tablet in front of a baby. Left half has images of gears and stuff, right half has images of people and faces. Does the baby look at the left or right half of the screen? This is actually pretty indicative of autism and easy to put into a foolproof app.
The linked github is recording a video of an older child's face while they look at a person who is wearing a camera or something, and judging whether or not they make proper eye contact. This is thematically similar but actually really different. Requires an older kid, both for the model and method, and is hard to actually use. Not that useful.
Intervening when still a baby is absolutely critical.
P.S., deciding which half of a tablet a baby is looking is MUCH MUCH easier than gaze tracking. Make the tablet screen bright white around the edges. Turn brightness up. Use off the shelf iris tracking software. Locate the reflection of the iPad in the baby's iris. Is it on the right half or left half of the iris? Adjust for their position in FOV and their face pose a bit and bam that's very accurate. Full, robust gaze tracking is a million times harder, believe me.
Thats a cool idea, thanks for sharing! It's cool to see other uses for a model I built for a completely different task.
Is there any research/papers on this type of autism diagnosis tools for babies?
To your last point, yes I agree. Even the task I setup the model for is relatively easy compared to proper gaze tracking, I just rely on large datasets.
I suppose you could do it in the way you say and then from that gather data to eventually build out another model.
I'll for sure look into this, appreciate the idea sharing!
Idk of any research, sorry, just going from memory from a few years ago. Feel free to lmk if you ever have any questions about mobile gaze tracking, I spent several years on it. Can you DM on here? Idk.
FYI: I got funding and gathered really big mobile phone gaze datasets and trained CNN models on them that got pretty accurate. Avg err below 1cm.
The whole thing worked like this:
Mech Turk workers got to play a mobile memory game. A number flashed on the screen at a random point for a second while the phone took a photo. Then they had to enter it. If they entered it correctly, I assumed they were looking at that point at the screen when the photo was taken and added it to the gaze dataset. Collecting clean data like this was very important for model accuracy. Data is long gone, unfortunately. Oh and the screen for the memory game was pure white, which was essential for a reason described below.
The CNN was a cascade of several models:
First, off the shelf stuff located the face in the image.
A crop of the face was fed to an iris location model we trained. This estimated eye location and size.
Crops of the eyes were fed into 2 more cycles of iris detection, taking a smaller crop and making a more accurate estimation until the irises were located and sized to within about 1 pixel. Imagine the enhance... enhance... trope.
Then crops of the super well centered and size-normalized irises, as well as a crop of the face, were fed together into a CNN, along with metadata about the phone type.
This CNN estimated gaze location using the labels from the memory game derived data.
This worked really well, usually, in lighting the model liked. Failed unpredictably in other lighting. We tried all kinds of pre-processing to address lighting but it was always the achilles heel.
To my shock, I eventually realized (too late) that the CNN was learning to find the phone's reflection in the irises, and estimating the phone position relative to the gaze direction using that. So localizing the irises extremely well was crucial. Letting it know what kind of phone it was looking at, and ergo how large the reflection should appear at a certain distance, was too.
Making a model that segments out the phone or tablet's reflection in the iris is just a very efficient shortcut to do what any actually good model will learn to do anyway, and it will remove all of the lighting variation. Its the way to make gaze tracking on mobile actually work reliably without infrared. Never had time to backtrack and do this because we ran out of funding.
The MOST IMPORTANT thing here is to control what is on the phone screen. If the screen is half black, or dim, or has random darkly colored blobs, it will screw with where the model things the phone screen reflection begins and ends. HOWEVER if your use case allows you to control what is on the screen so it always has for instance a bright white border, your problem is 10x easier. The baby autism screener would let you control that for instance.
But anyway, like I said, to make something that just determines if the baby is looking on one side of the screen or the other, you could do the following:
1. Take maybe 1000 photos of a sampling of people watching a white tablet screen moving around in front of their face.
2. Annotate the photos by labeling visible corners of the reflection of the tablet screen in their irises
3. Make simple CNN to place these
If you can also make a model that locates the irises extremely well, like to within 1px, then making the gaze estimate becomes sort of trivial with that plus the tablet-reflection-in-iris finder. And I promise the iris location model is easy. We trained on about 3000-4000 images of very well labeled irises (with circle drawn over them for the label) with a simple CNN and got really great sub-pixel accuracy in 2018. That plus some smoothing across camera frames was more than enough.
Anyway, hope some of this helps. I know you aren't doing fine-grained gaze tracking like this but maybe something in there is useful.
wow, this is great! You can't DM but my email is in my blog post, in the footnotes.
Do you remember the cost of Mech Turk? It was something I wanted to use for EyesOff but never could get around the cost aspect.
I need some time to process everything you said, but the EyesOff model has pretty low accuracy at the moment. I'm sure some of these tidbits of info could help to improve the model, although my data is pretty messy in comparison. I had thought of doing more gaze tracking work for my model, but at long ranges it just breaks down completely (in my experience, happy to stand corrected if you're worked on that too).
Regarding the baby screener, I see how this approach could be very useful. If I get the time, I'll look into it a bit more and see what I can come up with. I'll let you know once I get round to it.
We paid something like $10 per hour and people loved our tasks. We paid a bit more to make sure our tasks were completed well. The main thing was just making the data collection app as efficient as possible. If you pay twice as much but collect 4x the data in the task, you doubled your efficiency.
Yeah I think its impossible to get good gaze accuracy without observing the device reflection in the eyes. And you will never, ever be able to deal well with edge cases like lighting, hair, glasses, asymmetrical faces, etc. There's just a fundamental information limit you can't overcome. Maybe you could get within 6 inches of accuracy? But mostly it would be learning face pose I assume. Trying to do gaze tracking with a webcam of someone 4 feet away and half offscreen just seems Sisyphean.
Is EyesOff really an important application? I'm not sure many people would want to drain their battery running it. Just a rhetorical question, I don't know.
With the baby autism screener its difficult part is the regulatory aspect. I might have some contacts in Mayo Clinic that would be interested in productizing something like this though, and could be asked about it.
If I were you, I would look at how to take a mobile photo of an iris, and artificially add the reflection of a phone screen to create a synthetic dataset (it won't look like a neat rectangle, more like a blurry fragment of one). Then train a CNN to predict the corners of the added reflection. And after that is solved, try the gaze tracking problem as an algebraic exercise. Like, think of the irises as 2 spherical mirrors. Assume their physical size. If you can locate the reflection of an object of known size in them, you should be able to work out the spatial relationships to figure out where the object being reflected is relative to the mirrors. This is hard, but is 10-100x easier than trying end-to-end gaze tracking with a single model. Also nobody in the world knows to do this, AFAIK.
ha, thats probably why I noticed the EyesOff accuracy drops so much at longer ranges, I suppose two models would do better but atm battery drain is a big issue.
I'm not sure if it's important or not, but the app comes from my own problems working in public so I'm happy to continue working on it. I do want to train and deploy an optimised model, something much smaller.
Sounds great, once a POC get's built I'll let you know and can see about the clinical side.
Thanks for the tips! I'll be sure to post something and reach out if I get round to implementing such a model.
The point is that Gaza receives a lot of attention because we (the West) are involved. And its especially emotional because our politicians have been corrupted and blackmailed into complicity by people like Miriam Adelson and Jeffrey Epstein, and are taking us along for the ride, which is extremely frustrating. On the other hand, I have nothing to do with events in Sudan that I'm aware of.
The IDF was formed from the merger of 3 terrorist organizations responsible for bombings and murders of Palestinian, British and even Jewish civilians. Not much has changed.
They were forced to disband, lay down their arms and merge into pre-existing IDF, specifically because the methods are not acceptable, some of them declined and had to be disbanded by force.
Did you know that Menachim Begin, the first prime minister of Israel, was the leader of Irgun?
'As head of the Irgun, he targeted the British in Palestine,[3] with a notable attack being the King David Hotel bombing. Later, the Irgun fought the Arabs during the 1947–48 civil war in Mandatory Palestine and, as its chief, Begin was described by the British government as the "leader of the notorious terrorist organisation".'
The targeting the British happened primarily during the holocaust when British were limiting Jewish migration (which has directly resulted in many many deaths).
The history is complex and it is very easy to find reasons to dismiss people who lived through very difficult times and had to make very difficult choices.
The first prime minister of Israel was David Ben-Gurion.
Targeting the British during the Mandate was an act of anti-colonization, not unlike other factions that wanted independence from the British at that time, and the Irgun did warn the British before hand but they did not evacuate the hotel which was the central offices of the British Mandate and the Headquarters of the British Army
There is nothing to whitewash about wars and war crimes, they happen in every war, and are inexcusable in every war, no matter the side.
For every war crime from Jewish side you can always point to a similar one from Arab side, this finger pointing is not useful for anything but spurring more hate and igniting the conflict over and over.
I disagree completely. Its important to condemn terrorism and genocide on both sides. I have no tolerance for that. Its absurd to try to convince me to ignore this, or to try and excuse the IDF because the other side is bad too. Hamas is bad. Nazis are bad. The IDF is bad. The Israeli state is bad. Israeli society is rotten. These are all true and useful statements. I don't understand why you would spend your time on the internet trying to argue against this.
"Hannah Arendt and Albert Einstein, in a letter to The New York Times in 1948, compared Irgun and its successor Herut party to "Nazi and Fascist parties" and described it as a "terrorist, right wing, chauvinist organization" "
Thanks so much for the kind words - its 100% o3-mini for clustering. I have zero editorial input as to what constitutes a cluster, what's "top" news, etc.
The one subtlety is setting up the LLM to understand whether a new story belongs in an existing cluster, or with > 1 neighbors, constitutes a new cluster. The challenge there is scoping the clustering window (hours of stories for consideration) and topic breadth to avoid creating Katamari-super-clusters that just end up with every story associated to them.
At this point I seem to have found a sweet spot re: the hours window, the frequency of processing, and the design of the prompt such that its working consistently.
Very few false positives in terms of spurious clusters being created, or potential clusters being missed.
Let's take this argument to its logical conclusion. If every single person on Earth moves to America as an H1-B we would get 100% of the Jensen Huangs!
Not rightwing. Israeli. Larry Ellison is the biggest donor to the IDF. Tiktok was banned in the US when the youth started to turn against Israel. This has everything to do with propping up Israel's cratering reputation with the youth.
The vote to ban Tiktok (for spurious "Chinese data theft" reasons) was pretty much the only bill to have bipartisan support in the last decade. Goes to show how deeply the lobbying influence is in US politics.
Have you read TikTok's ToS? The claims are not spurious. The amount of data collected for each user is extraordinary. It is enlightening that we are still even talking about it.
The TT "ban" went into effect in January, and wasn't enforced. Their data practices don't seem all that dangerous if they've been able to continue them unemcumbered for more than a year after the bill went to a vote.
I mean the reality is that going forward it simply won't be permissable to let foreign governments freely influence your youth with whatever content algorithm they please. It's even open knowledge that China itself doesn't allow TikTok to influence their own children like they allow it to influence American ones. As influencing physchology and personal Ai algorithms get more advanced you simply can't let it be controlled by foreign, hostile powers. This was inevitable, lobbying aside, as long as TikTok has suge huge market share especially with children.
Netanyahu is bragging about how they bought TikTok lol. This isn't about China, and it definitely isn't about keeping foreign governments from influencing our youth. Its the opposite! China was the pretext to get you to swallow another foreign government controlling TikTok.
What exactly were people banned for? What is a "lie"? If I say I doubt the vaccine will prevent transmission, or I suspect the virus leaked from the coronavirus lab down the street, would I have been banned? Honest question, I don't know precisely why these people were banned.
Am I the only one here who is skeptical that the MRNA shots have a positive risk-reward tradeoff for healthy people?
My job is to rush out complicated things, and biology is much more complex than even hardware. I know how fallible experts are. I know what hysteria and pressure driven deadlines do to build quality. And I know that you can't really test long term effects of biological products without long term human trials.
I think everyone here knows these things too, but most have a tribal political reaction to trust these shots because you dislike the people questioning them.
Do you have the same reservations for the other 99.9% of drugs that haven't gone through "long term human clinical trials?" How long would a trial need to be to assuage your concerns? How large would it need to be?
Because even the one "significant" adverse effect we see with COVID vaccines (myo- or pericarditis in young men) is so uncommon that it would've required trials several orders of magnitude larger than the largest trials ever conducted to detect.
I personally don't mind people asking questions of them, that's the whole point of science!
What's frustrating is when people ask questions that 1) are easily answered by the reams of publicly available literature on the topic [and are truly difficult to answer in 240 character snippets], and 2) they never even thought to ask of the various other compounds they're ingesting which are generally far less well-understood than the COVID vaccines.
It's almost like it's not an organic and earnest curiosity...
Anyway to your direct question: no obviously you're not the only one. But according to all available data the calculus very strongly contradicts your intuitions. We make decisions with incomplete data every day and we don't generally call "filling in the blanks with whatever I need to in order to reach a contrary conclusion" a form of wisdom.
