I think i found answers to my question.

In Vuforia SDK v7.1.34 release notes (https://library.vuforia.com/articles/Release_Notes/Vuforia-SDK-Release-Notes.html) are many known issues for model targets. For example there is a known issue for cluttered backgrounds. So i think it will be solved in later releases.

And i saw Model Targets with Advanced Recognition will be part of Vuforia 7.3, which is currently in beta development.

So I think this feature is under development, and I hope it will be released in the near future.

I have only one question to VuforiaDevs. When will  these features (especially false positives on cluttered backgrounds) to HoloLens release?

Best wishes

Benjamin

Hi,

I'm trying to recognize a real car (a FiatPunto).

The recognition is quite good, but there are a many false positive recognitions.

I'm using diferent colors in the 3D model for different materials and the scale is also good.

I'm attached an example for the false recognition. I tried with different Guide views, but the result was same, it recognised a car on my table every time (and on my keyboard every second time).

What else can I do to reduce the false positive recognition?

Will there be a better solution in the near future?

Technical details:

- Unity version : 2017.3.1p1 ( i would like to use this model recognition feature on hololens but currently im testing on an android device first)

- Vuforia version : 7.0.47

- Device : Pixel 2 XL

Thanks,

Benjamin

Hey, Join the Club,

Even i had this problem earlier, the following may help it to significant extent:

• use different colors in the 3D model for different materials (e.g. plastic, chrome, metal, glass). The specific color is not relevant (i.e. the 3D model's colors do not need to exactly match the Model Target's), just the differences in colors.
• Check your scale
• Try with different angled Guide views

Even trying all the above things, you may encounter false positives , with the things that are almost alike your target objects. Hope this helps.

Regards,

saikrian.D 

Almost the exact same post I did 3 days ago...

You're not alone mate ;)

I'm hopping I did something wrong but I'm not so optimistic anymore. A quick word from the team would be really great :D

Cheers,

Pierre