[Enhancement] Optionally return InputImage to be further used with other ML Kit libraries

[iklochko]

Hello.

Thank you for developing and supporting this amazing library! I use it to scan barcodes, including ITF, which is reported to be scanned unstable, and it really is. Luckily, in my case i have this barcode duplicated with symbols under it. I've managed to succeed obtaining good results in following way:

1. I receive the Barcode from mobile_scanner. returnImage is not working on Android, and it's crashing the app on iOS sometimes (added some details here)
2. Then, I have to stop the scanner and use another lib to take a photo.
3. Then I use google_mlkit_text_recognition to match Barcode with text recognized. I have to save photo to directory to convert in to InputImage used by ML Kit.
4. This in general works fine, but there are some performance issues on old devices or many simulate recognitions.

So, I wonder, if mobile_scanner also use ML Kit under the hood and it also obtains InputImage to recognize barcode, will you please consider a possibility to add an option to return InputImage with scan results?

Best regards