Image Completion

Developed deep learning models (using PyTorch framework) to predict the complete human face when it is partially covered by COVID-19 masks.
DCGANs outperformed CNN AutoEncoders and Visual Transformers in reconstructing the missing facial area.

Report and Results