Dr. Safi Ullah
During my MSCS, I worked on Self Correcting Images. In PhD I worked on Single Image Super-Resolution. In PhD, we specifically addressed zero padding issue in image super-resolution. we proposed an adaptive compensation technique for zero padding in the Residual Networks. I also worked on optimal samples selection from a huge dataset to constitute a suitable training set (Note that training set is one of the major potential cause for poor network generalization). Currently, I am working on medical image segmentation.
My research intrest includes Single Image Super Resolution, Training set selection, Fake news detection, Generative adversarial networks, Federated Learning, Explainable AI.
- PhD, Other
- MSCS,National University of Computer And Emerging Sciences, Islamabad
- MIT,Kust
- B.Sc (2 Years),Gomal University
- S.Sc,Bise, Bannu
- S.Sc,Bise, Bannu
- Member DRRC of IT department
- Department of IT focal person for NCEAC
-
1. Safi Ullah, Seong-Ho Song “Design of compensation algorithms for zero padding and its application to a patch based deep neural network” Peerj, August 2024 DOI:
-
2. Safi Ullah, Seong-Ho Song “SRResNet Performance Enhancement Using Patch Inputs and Partial Convolution-Based Padding” Cmc-Computers Materials & Continua, June 2022 DOI:
| Date | Title | Agency/Organization | Amount | Status |
|---|---|---|---|---|
| 2021-02-01 | SRResNet Performance Enhancement Using Patch Inputs and Partial Convolution-Based Padding | Hallym University Research Fund HRF-202104-004 | $10,000 | Completed |
| 2022-06-01 | Design of compensation algorithms for zero padding and its application to a patch based deep neural network | Hallym University Research Fund HRF-202208-004 | $10,000 | Completed |