Do I need to do preprocessing ?

[Sharkoon1]

Hi,

I am new to Firebase and the MLKit. I was wondering if any preprocessing is required or if this is already being done by the plugin ?

I want to recognize urls out of an image from a browser. The mlkit plugin has problems with readings f.e equal signs or hyphens.

The original link: https://www.youtube.com/watch?v=F5KgsubkMCA

After Firebase OCR: https://www.youtube.com/watch?v-F5KgsubkMCA

As you can see "v=" is later "v-", sometimes also ":" are missing.

Is there a chance that you could implement some preprocessing ?

[azihsoyn]

Hi, there!

Firebase MLKit piovides preprocessed model. So, you don't need to preprocess.

If you want to customize detected result, Firebase MLKit supports custom model detection, or cloud base detection.

Currently, flutter_mlkit is not supporting both feature.

Custom model detection is under development, but it will take a just time.

Thanks.