Person Detection Update: A New Experiment for Premium Features

you are correct with xnor
I am thinking Wyze home brew codes build into firmware.
Fact is detection alogrithm is taken out of .156 firmware. I would think .111 still have embedded detectionusing Wyze home brew.

No they never had their own “home brew” local / embedded / edge person detection, as far as anyone knows. It was Xnor and then cloud.

2 Likes

Cloud (Pilot) PD is in x.156.

1 Like

@UserCustomerGwen I’m sure someone has already mentioned it but why give if us the ability to build polygon based detection areas rather than rectangular ones? If we could have better defined areas that would be less motion events which means less video to run person detection on. Also means less useless alerts for us.

5 Likes

It’s most likely built into the camera firmware which can’t be changed.

1 Like

Agree on all points! :+1:

It’s the #5 top request - you may vote for it here - it may do some good.

2 Likes

Better still zoneS should be selected from a grid. Then multiple zones can be made of individual pixels or squares of multiple pixels. 4x4, 8x8 etc
Just as Reolink have done

1 Like

It wonzt be easy to store a few million detection setting in server and every setting is different.

This can only be done if the cam has the local processing power.

Wyze cam does not have enoigh rom, ram and cpu power to do the task.

Agreed! :+1: I like and have used grid zone motion detection. It seems the simplest and most flexible system to create any shape. Obviously the finer the grid the better.

But apparently a lot of people like multiple polygons or multi-point polygons. So, I’ll endorse anything better than one rectangle at this point.

I presume the current motion detection zone filtering is done at the cam level, otherwise the entire feed would have to constantly be streaming to the Wyze cloud servers for processing, yes?

So, this is something I’d really like to know the answer to - just how much more processing would a grid need over the current single adjustable polygon?

Basically, would it fit in the space and processing time freed up by xnor code removal?

My xnor firmware cams behave an awful lot like the current firmware cams, so if the xnor code is just being ignored, then what is using that space and CPU time in the current firmware? Anything useful to the consumer? Could it be used for better detection zones and momentary event filtering?

=

It seems to me that CAMPlus “solves” a bunch Wishlist items so that Wyze doesn’t have much incentive to invest resources in offering them for free - unless they want happier existing customers… :wink:

  1. FF/RW in playback - get CAMPlus to save long clips with no cooldown

  2. Greater control over detection zone - get CAMPlus if you want to ignore extraneous events

  3. Ignore motion events under X seconds - get CAMPlus if you want to ignore extraneous events

  4. Direct access to SD card with D/L ability - get CAMPlus if you want long clips and no cooldown. You can download them from the cloud if you want to keep them.

  5. Mark events in playback - get CAMPlus if you want long clips and no cooldown

  6. Network attached storage - get CAMPlus to store want long clips with no cooldown. You can download them from the cloud if you want to keep them.

  7. Loss of connection notifications - future premium service like emergency button? Probably not enough demad?

=

I wonder if there is hope for any of these?

  1. Multiple user permission levels

  2. Web browser interface

  3. App dark theme/mode

  4. Tablet app landscape mode

  5. SmartThings integration

3 Likes

Good post. Wise even.

2 Likes

For some reason that struck my funny bone square on this morning! Thanks I truly needed a reason to smile this morning! :slightly_smiling_face::slightly_smiling_face::slightly_smiling_face:

1 Like

I want to start by saying I love you guys. But I am also worried about the person detection, I do not want CMC as someone coud just notice the camera, and snatch it, so the camera will never upload the video to the server since motion was still ongoing. If fee based, can’t we just pay for person detection, not CMC?
Also, how about you have x hours of free person detection per month, you pay for additional hours? Quite a few, like me, only activate motion detection and notifications when away from home, it does not have to run 24/7, and does not have to run on every motion detected (if the camera is on, but we are home)

1 Like

WYZE changes the firmware continually.

3 Likes

@Megamind @xmick

I tried Android 9: rockpi-4b-rk3399-android9-20200522_1559_2f865d4-gpt.zip from the site xmick mentioned. I successufully donwload/eteched the image, but as best as I can tell, I hadn’t realized that Rock Pi is sufficiently different from Raspberry Pi, such that the Andriod 9 image will not boot on my Raspbery Pi 4B (2GB).
End of that experiment for me: (something for thers to try)

I was successful running tinyCam for Wyze, Foscam and Tenvis cameras (Yi cameras do NORT work) on BlueStacks, but I don’t want to keep a Windows systems powered up for this. I’d much prefer a to have a Raspberry Pi powered on 24/7… so I look at Home Assistant on RPI and see where that takes me.

1 Like

I’m sorry, I’m a bit confused here so I just need some clarity.

  1. I’m confused about $0 optional pay what you want and Free for those who “signed up” with Wyze before Nov. 2019 Person Detection. On one hand I’m seeing that “Free” Person Detection is going to go away in Sep, but that people with accounts created before Nov 2019 will still be able to get Person Detection for free.

I take this as, moving forward and for people who created accounts after Nov. 2019, CMC Plus will be the only way to get Person Detection, since it is cloud based. People who had their accounts before Nov. 2019 will get to keep whatever form of Person Detection we have right now. Am I right with this?

  1. How do I know if I am considered someone who signed up before Nov. 2019. I know I’ve been purchasing Wyze Cameras since they were released, starting at V1s. I know that I’ve had an account since then. I know the website went through a change and we had to change passwords. I have been posting on the forums since 2018 or so, I think. But there’s nothing on my profile page showing creation date, etc.

  2. I purchased MORE Wyze cams, due to person detection so I am actually one of those people who bought into the whole, “free person detection” promise and it is important to have it on certain cameras.
    I am considering doing CMC on ONE camera to trial it, but it is very important that if I don’t want to keep paying for it that the camera will go back to doing whatever version of Person Detection it is right now (I’m assuming I’ll be grandfathered in). You’ve mentioned that this ‘grandfathering’ is account vs. camera based so am I correct in assuming that if I enable CMC Plus and then disable it, the camera will go back to it’s normal method of Person Detection vs. not do it at all?

Thanks.

2 Likes

You should be fine, although only the magic Wyze server knows for sure.

1 Like

I was never on the earlier version of the forum, but I definitely purchased Wyze cams long before visiting the forum. So is it the forum signup date or the date you first created a Wyze account in an app?

The whole thing gives me a headache now, when all I wanted was free working PD to continue…

You look safe either way. Congratulations?

Also, it’s got to be the account creation date, not forum date, that matters, I would imagine.

1 Like

It’s just one thing. Pay what you want if you had an account before Nov 2019.
People are calling this free but technically it isn’t.

Both forms of Person Detection are the same cloud based program.
The paid subscription version come bundled with Cam Plus (CMC). You have no choice on the bundle.
The older customers have the option of stand alone Person Detection without CMC.

1 Like

So how and where exactly would I pay what I want?

Thanks