The Pet Option is back in the filter options for me today too. It still doesn’t show the “pet” tags on the events though. It also isn’t showing under "Account - services - Cam Plus - [cam name] - “Pet Detection” as an option I can turn on or off like the other detections.
It is interesting that it is showing back up without any kind of app update. It must not be part of the app itself then, but basically HTML which they can change without a new app version release.
I live on a canyon in Southern California, and as has been witnessed over the last several years, we have more than our share of wildfires. The Wyze cams currently alert us to in-house smoke alarms and CO2 alarms, but many of our problems occur with fires that approach from outside the home. It would be great to have a feature where the camera could detect the heat signature or flame pattern of an approaching wildfire. These things move incredibly fast and any time saved would be invaluable. Thanks.
Most needed motion capability - should pay for itself in lower data volume.
(Now running 4 Wyze PLUS, Ring and Nest cameras in side by side tests)
A- add PLUS ability to identify moving grass or branch back and forth motion. Option to not report OR
B- Ability to add Motion-Do Not Report pentagon area so locations of frequent wind blown plants are not reported , sent to Wyze, when Report Motion is turned on.
C- This will 1- allow invasive animals to be detected and
2- greatly lower false plant motion data volume being sent/ processed by Wyze.
4- Add a must have PLUS capability no < $300 camera has.
Can you elaborate on how this is already different than what is already offered by the Cam Plus subscription together with detection zones?
We can set up detection zones to ignore or only look at certain areas. Wyze is also updating the AI to recognize these zones and only analyze the detection area (not the entire video). These detection zones can be set up by choosing individual little grid areas (it’s divided into 10X16 squares on the V3 cams, so 160 different squares, to customize our preferred detection zone).
We can choose specific detection options (humans, animals, etc) and it will ignore all motion from grass and trees and only give reports about humans or animals in the recording in the detection zones we choose.
It seems to me that with Cam Plus AI detection, we are able to set up detection zones, ignore grass and trees, still get alerts for humans and animals, etc. I use this all over my property. It’s especially critical at the front of my house. My entire front yard was torn out and replaced as a flower garden, so there is literally CONSTANT movement triggering motion detection, so I got Cam Plus, and set up detection zones and now I only get alerts for person and animal detection. Rarely to never do plants trigger alerts for me anymore, despite them covering the whole front of my house. I get told when people come to my house, when my cats come home, when Racoons pass by, etc. To me, it seems Cam Plus AI works great for this as it is…
That is why I am asking for some clarification (so that I understand)…Is Cam Plus not doing the above already? Or were you suggesting that you want all of that to be working locally on the cams without a subscription? That I can certainly understand. I would love some edge-AI improvements too, though I don’t think local edge solutions are WyzeShawn’s expertise, though I saw they posted jobs for local edge programming last year, so hopefully someone will look into that because filtering out plant movement without a subscription would be incredibly valuable for lots of people!
Hey @WyzeShawn you said here that you were looking for Wyze Bear videos to help train the AI and I wanted to let you know that @tboilard just posted a couple of bear videos on Wyze Cams and gave permission for you guys to have/use a copy of the videos:
Apparently it was too late to submit the videos through the normal route, but I thought you might still want a copy.
Hey guys, looks like this might’ve finally moved out of internal testing! I am showing access to enable facial recognition again and sort by faces (you have to manually add facial recognition again by going into Cam Plus, select the camera and click on accept for facial recognition to re-enable it now that it’s back).
All my previous facial profiles from Version 1 of this feature are gone, and I haven’t had any option yet to set up new ones with recognized faces. I hope someone else will update us when they are able to add facial profiles and confirm it’s working in events again. @spamoni4 I hope you’ll test this out too.
I could’ve sworn you had registered before with all your testing help on the other AI stuff! And that once the other AI stuff went to everyone the registration was no longer needed especially since all this stuff has been getting discussed in the “News From Wyze” section for months now instead of the “Beta” section. My apologies for being wrong. Go apply if you’re interested.
Yes and no. Of course Wyze has and uses an API, but they haven’t made anything public or official yet for tinkerers to use officially to integrate Wyze stuff into other things.
Having said that, there are options on GitHub and other places that people have been using to access and create their own API integrations into various other things. I use one to link my Wyze devices to Home Assistant for example. Some people use the RTSP firmware for the V2’s to stream their video to various other programs. and Wyze is also about to release their RTSP for the V3’s. You could then stream your own video and audio through some other service like AWS or some localized AI to try to transcribe camera feeds.
That’s a little above my expertise, but it is definitely something that could be done. In theory, if you just streamed all audio to be processed by a dictation program, you might be able to do all that virtually for free. Still, Wyze’s audio capture isn’t always the most clear, so hopefully future cameras will have improved microphones to be able to do this. I would just think that Wyze would be interested in having their AI do it all while it’s analyzing all the captured video and audio already…why not add this to it to make it even more useful?