Detection sensitivity and notifications

[STTH]

The detection being ultra sensitive is a huge benefit. Only as long as it can be managed. One nice feature is the ability to reduce the number of notifications based on the type of detection (person, animal, etc...).

Is it possible to likewise reduce the number of recordings based on the type of detection? I might want to capture all the persons, but not the cars. Or I might want the persons and cars, but not the other types of motions.

If it is not possible to avoid recording all types of detections, it is not a big issue as long as we have enough storage and bandwidth (and let's not talk about battery life obviously...). What Arlo developers could then do, is simply add an option to automatically delete the recordings based on the type of detection selected, right after the recording has been completed. This would be extremely easy to implement, all on the software end.

The idea, on one post, that users should refrain from pointing the cameras too high so as to pick up every car is somewhat absurd: it all depends what you are trying to accomplish. If you're dealing with security and have one camera for close ups near your front door with no view on the street, there is still plenty of benefit from having a second camera pointing toward your front yard in hope to record license plates of passing cars.

Especially when Arlo advertises how it can zoom in on plates due to the high resolution of the 4K cameras, what good is it to be alerted that someone intruded your home if you don't have another camera assisting to a reasonable degree in trying to see if you can also catch the guy's vehicle (and associated identification through the license plate)? These cameras are increasingly sold as multi-packs specifically to get multiple angles; point being, you will have some cameras that need to point up a bit.

But the whole idea is to intelligently filter the content of what is being captured. The whole difficulty is how to implement a sophisticated algorithm to identify vehicles, people and animals; even packages... and guess what: Arlo already does that! So at this point it's just a matter of enabling the user to control which should be active or not. There is really not much to it.

What Arlo can also do to give the user a measure of control on the sensitivity of the detections, is to for instance implement an invisible bluriness mask, with a degree that is user defined, and which will reduce the sensitivity simply because it will mimic an inferior resolution just for the task of detection.

So there really are options out there, and the funny thing is... they are for the most part software-based and relatively easy to implement.

Bottom line is this: Is there any such thing already in place, and if not, what are Arlo's plans to that end?

[STTH]

When creating a rule within the detection Modes (to trigger video recordings), is it possible to define the sensitivity per type of motion (person, animal, vehicle, other)? For instance, have the person detection sensitivity set to 100% but the vehicle detection sensitivity set to 50%.