Help in Understanding why the model fails/mispredicts for certain image's orientation detection (english language based ID Cards).
#2
by khushalR - opened
Need help in understanding what features contributes towards the model's prediction or misprediction in different cases. And are there any limitations where the model's mispredicts so we can effectively capture and improve that data capturing mechanism to mitigate such issues? Facing problems with Indian ID Cards (Aadhaar, Pan, Passports, VoterIDs, etc).