Vehicle events being falsely labeled as people

So for the last two weeks or so my v3 looking out to my front porch and yard seems to be falsely labeling vehicles as people .

I submit the correct feedback but find it a bit strange , it’s been notifying me of person events over vehicle events for over 2 weeks . About 10-15 false positives is what I think it has labeled

Funny enough , for the last 2 night when i walk past this exact v3 during the late afternoon it’ll label me as just basic motion. As with the AI it can never be perfect but it makes you kind of distrust the AI in total…

1 Like

It is possible that it is not identifying the vehicle as the person, but something else in view that has the rough shape of a person, including possibly a shadow cast by the headlights or something else. I have seen events where something else triggers the event, including cars, and getting person identification, then I tell them to move something and it fixes the false detection. For example, one person who was getting false person detection, I was confident it was their chair causing this (because I had a similar situation with my eufy cams falsely identifying chairs as “persons” with their AI. So I told them to change the chair (move it, rotate it, throw something on it, whatever), and as soon as they did, the false person detections went away. In other cases I’ve seen certain objects in people’s yards be identified as a person, sometimes even a shadow that occasionally has a person-like shape sometimes.

In other words, it may be that the vehicles are triggering an event and something else in your yard or your view has a rough person like shape to the AI, or is commonly associated with so many person submissions that it thinks it’s related to a person somehow.

Most of those things will be resolved over time with enough training data, but my bet is something in view other than the cars causing this. I’ve seen it many times.

It can be hard for us to know for sure what is being misidentified…but sometime this year Wyze is releasing a new feature that will tell us exactly what object was identified as what, so we’ll easily be able to solve problems just like this! I look forward to that feature so we can figure out the issue and resolve it quickly and train the AI with very good examples.

5 Likes

Good morning Carver

How do you have the answer for everything ?!?!

Yea so my v3 is mounted to my front window overlooking by front porch and front yard that ny street is also in frame so cars drive up and down the road all the time…

So the camera is over looking my porch and I have a chair right in front of it that I often times sit on and any little movement will label it as a person and I submit the feedback as a person all the time. So I assume that when the vehicle was motion tagged it only tagged it at the end when it was going out of frame , i think it automatically thought it was a person bc when i sit in the chair in frame and i move my head or any little movement will be labeled as a person so I assume the AI has been reviewing the events and thinking that way for some time

Thank you for the answer my friend

3 Likes

Thanks buddy, and you’re welcome. :slight_smile: I’m getting ready to go to bed now (7am my time)

For now, just submit enough video examples for the AI to have plenty of examples when the identification is correct or not. Especially submit as many videos as possible when it falsely identifies a person, and make sure to uncheck the person tag so that it tells the AI it was wrong and consider what is different from the correct ones. It may take a couple/few months of new training data submissions, but I have had some similar situations that I absolutely saw a huge difference over time of me submitting corrections and also sometimes correct detection videos.

I had a camera that nearly 100% of the time always detected my black cat as a person at night when he walked up my driveway, Every time they submitted new data I saw a big difference. It happened less and less often until it basically never mis-identified my cat on that cam anymore. Basically from 100% of the time to almost 0%…but it took a while. It wasn’t a fast thing. Together with other similar examples from others it does learn over time though. I’ve been using the AI features since really early on and I have absolutely seen drastic improvements over time. It will only get better.

Also, in the short term, you can see if a slight repositioning of the chair or other things that could be affecting it helps to reduce the frequency until the AI improves more for that situation.

When you get up you got some explaining to do… Eufy cams!?! Traitor!

I never went to bed yet :sob: My kids needed some help…

Sure. I have 3 eufy cams and over 2 dozen Wyze cams. Wyze is definitely my primary tech company, but I do believe in some degree of having fail-safes and redundancies. If one goes down for some reason, the other catches everything, and it has been awesome to make comparisons against one of Wyze’s main competitors. I can also share my experience of why I still prefer Wyze over eufy with actual credible examples and experience. I heard a lot of people saying how awesome they were, so thought I’d try a couple out.

I also got my eufy cams a couple of years ago so I could get “Crying detection” Notifications, which Wasn’t possible with Wyze at the time, though Wyze has started working on Crying detection with their AI, they still don’t do notifications for it.

I also tried some Blink cams, but I literally gave that junk away to a friend who was desperate for security cams :rofl: I also have some slight experience with Arlo, and Ring. But I don’t really like them for various reasons.

But the main reason I have a couple options from another company is simply for redundancy/backup situations. Having said that, my eufy cams also give me 2K, and Wyze doesn’t currently offer that, though VP of Product said they’ll have some higher res cams at some point in the future.

I am not ashamed of having a few eufy cams. I don’t prefer them over Wyze at all. I wouldn’t switch to them. I’ll rarely get anything from them, but I LOVE that they cause a lot of competition. Competition is great for us users. :slight_smile:

1 Like

Fair enough, I got 2 ring doorbells myself.

Interesting that you think the billing is garbage, I was about to ask if you had tried those. They are obv a direct competitior to Wyze, (amazon trying to get their employees back lol), and they are backed by amazon so have a lot of funding and the ability to operate on a loss if needed.

I was worried they may be able to beat Wyze, but sounds like no?

1 Like

To each their own. they’re not as good as Wyze by any standard, with less features, etc. It just didn’t make sense to me. Sometimes you can get them cheaper though, especially around Black Friday and APD when you can add them on to a purchase of an echo show for just $5 more. That’s how I got a bunch of them. I preferred having eufy as my backup instead of Blink.

2 Likes

I confess I only skimmed the thread, but did you cover how the very latest firmware is supposed to resolve this longstanding issue by alerting only when the person is also in the moving area?

I did not. I was focussing mostly on how some things in the frame may be identified as a person when they are inanimate.

2 Likes

I appreciate your feedback, but I am experiencing the same issue where every car or truck that drives in front of my house generates a person detection notification. It’s basically useless now, I get so many incorrect notifications.

Side note, I’ve been using person detection since the V2 came out, so I’m not new to this feature.

Considering how this just started happening over the last month, it’s fairly clear the AI has been updated and it’s failing for me.

If you post an example video we might be able to give suggestions as to what might be getting misidentified in your videos (the car is probably the event trigger).

Have you tried blocking out the road with a detection zone? In that case, the AI should ignore the cars now. If you are using a detection zone, you could post a screenshot and get some suggestions.

Also, totally understandable if you decide not to post a video example or screenshot for privacy reasons. :+1: Just offering to see if we can help you reduce the issues you’re experiencing and see if any of us can see something we’ve seen similarly be misidentified in events and if there is any adjusting that can help.

Otherwise, you can keep submitting corrections to the events and this should help in the long run (not right away).

FYI, I just got a notice that the monthly AI update just dropped. While not drastic the following are supposed to be the measurable improvements this month:

Even if the improvement is just 1.5%-5% every month, after a few months of compounding improvement it is drastic progress…maybe not noticeable in a single month, but usually noticeable after a few months in a row.

2 Likes

That it did, hopefully some will see a bit of improvement.

4 Likes

I’ll go find some video examples tomorrow, but here’s the problem I see with your response. I ordered my V3’s via the waitlist when they first came out, and I haven’t moved them since I put them up. So Your approach of “we’ll help you figure out what’s wrong with your placement” isn’t what I need. They’ve been working fine up until recently, then I started getting all these bad notifications starting about a month ago (maybe a little more). So it’s not a placement problem. It’s an AI issue.

And to your question about removing the street with detection zone exclusions, the answer is no I can’t do that. Part of the reason I have these cams is to report when there are people near my car in front of my house. Besides, that’s just avoiding the real issue of inaccurate AI detection issues.

I wish I could report that my issues were happening at night, but they are all during the day in full bright daylight, so sounds like these updates won’t help my issues. Thanks for posting this though.

You are not alone. I recently did a 10 day log of AI tagged events. The results did not instill confidence. My worst 3 V3 cams were the three pointed at the street with PD only. Cars are constantly being PD tagged, mostly in daylight. This has been an ongoing shortcoming of the AI for some time.

One of the features of the Events page is that it will show you the snapshot thumbnail of the actual object that was tagged in the video:

Then, when you load that video to view, it will again show you that snapshot in the larger viewer before starting the video. It is there that you will be able to see what object caught the tag:

For example this vehicle being tagged as a person:

If you watch as I click the video, that same thumbnail appears prior to the full video loading:

Here are 10 videos of cars being tagged as people just from the last two days.




















Perfect examples! Thanks for going to the effort to post these. I know it’s a hassle.

2 Likes

That is a misunderstanding. The green tag is ONLY identifying motion (something that changed between 1 frame and another, not what the AI actually detected. The AI is considering each frame individually at the moment and the green tag is meaningless to it.

The green tag is 100% decided by the local Camera processing motion. The AI does not take this into account at all. AI detection is totally separate.

In your case above, The car is the trigger for the video, but more than likely it is another unmoving object being identified as a person. We can reasonably conclude it is not shadows (at least not most of the time) since many videos were not shadow related. My first guess would be one or more of the mail boxes being similar to the shape of a person at a distance. I have seen fire hydrants and house lamps and such be labeled as a person before. Possibly some other weird things like this:

image
image
image

All of which have a similar shape/outline to a human at a distance, or sometimes when there is over-exposure.

Don’t get me wrong. I am very supportive of the AI progressing to resolve that and come to learn the difference.

However, if the AI was actually identifying cars as people, then I would also always have cars identified as people, and I do not, and neither do the majority of people. Cars are rarely identified as people, and thus we can safely believe that it is not the cars themselves that are actually being identified as a person in these cases and that something else in your particular environment is triggering this mis-detection. There are countless potential suspects I found even just from brief inspection of the environment.

There are some things you can do to figure it out though. For example, just for testing sake (you can undo this later), try using a detection zone and blocking out a couple of specific areas that have a vague human-esque shape…anything that could appears to have a head on top of something else, or a head with arms or legs, or sitting showing just a body with a head, round or protruding in some way, whether or not a neck is apparent, even allowing for potential head differences due to hats or whatever. Block out mailboxes, stumps, anything with a roundish top, most of the tree branches, etc. Try lots of different things, maybe all together to start with and then just 1 or a few at a time. THEN see the difference. The AI has recently been updated to to respect most of the detection zones and ignore areas that are blocked out. Thus if you suddenly have fewer person detections when events trigger (from cars), you will know that something (or multiple things) in those blocked out zones are the primary offender giving you false person detections. Maybe try blocking out nearly everything except a piece of the road where it will still see a full size of a car, but not much else, so all your camera sees are the cars (make sure the mailboxes are blocked out, etc), then see that your person notifications decrease drastically because all it ever checks is the cars, and the cars aren’t what it thinks is a person…the AI is still seeing the cars, but it won’t be saying that they are persons. Then you can basically rule out the cars being the main culprit and expand your detection zone until you see you start getting false identifications again. Then you’ll know something in that area is actual culprit and that the cars are just the event trigger.

Having said all that, if you don’t want to go through all that effort, you can just wait. Either the AI will improve organically soon enough as it evolves with enough data, or by the end of the year they said they are going to release new features that will show us exactly what is being identified as what by the AI. Just today Jimmy confirmed that the AI team is working on having AI tagging in the video (just like you thought these events were showing with the Green tag, but were only tagging motion, not what the AI actually identified), AND he said they are ALSO working on having the notifications actually show us a thumbnail of the identified detection instead of showing us the entire frame. So by the end of the year we should be able to accurately see what your false identifications are, but I can nearly guarantee that the majority of them having nothing to do with the cars/vehicles…those are just the motion triggers that initiate the event.

Either way, Wyze is working on a solution to help you better identify what the real issue is.

image

1 Like

This is helpful. Thanks for the lengthy post. I have to question your comment that the AI respects “most” detection zone…

You need to explain that in more detail.