EMERGENT: barcodE iMage EnhanceR usinG dEep Neural neTworks

Authors:

• Aleksej Gaj
• Štěpán Studenovský

Task:

build and implement NN that will preprocess photos of damaged barcodes and enhance them to be readable by common methods.

photo of damaged barcode -> (NN) -> enhanced photo -> (existing barcode reader) -> the code

In case of low success redesign the barcode scanner using NN.

photo of damaged barcode -> (NN) -> the code

Notes

"Damaged barcode" = poorly readable barcode due to:

• dirt
• water droplets
• glossy surface
• curvy surface (ex. rounded bottle, bent package ...)
• blurry photo
• etc.

...was here:

Reminder

• Do not miss deadline for uploading your implementation and documentation of your final project.
• Include working demo in Colab, in which you clearly demonstrate your results. Please refer to the Implementation section on the labs website.
• Do not upload huge datasets to GitHub repository.