figured if we all upload clips of Wyze’s ML / object recognition failures to detect correctly, their engineering team can benefit from a concentrate of these types of clips, and use them to reproduce the failure, thereafter being able to correct it - and we all benefit.
Starting it since I found nowhere under “Support” to communicate such issue to support, let alone engineering.
Welcome to the forums! When you are looking at an event in app, there is a link at the bottom to submit the video to help teach the AI.
and here’s my first - movement detection failure spectacular …
I mean literally yes, it’s a movement, but we all know we’re seeking 3 dimensional movements (i.e. human-relatable events like theft, proximity, events near by involving objects) - in short - I don’t think identifying a shadow of branch waving in the wind applies.
This movement happens near my house as the wind blows, and even at min sensitivity it’s reported as movement and recorded constantly which is annoying and impossible to get rid of lest I wish to get rid of all movement detection altogether - which is clearly undesirable as thieves don’t ring doorbells…
here’s link to the clip - https://1drv.ms/v/s!AnyZXoU3lLlUgppTeoXyFJlE7XbBVQ?e=twiA7w
