Fahad Shamshad
fahad.shamshad3 <at> gmail.com
Update: I am now a PhD student at MBZUAI, focusing on the safety aspects of Large Multimodal Models. For potential academic collaborations or inquiries, please do not hesitate to email me or message me on Linkedin. Stay tuned, I will update this page soon!
I am working as a researcher at Computer Vision Department at MBZUAI, Abu Dhabi. Before that I worked as a machine learning research engineer at OMNO.ai, a Pakistan based technology startup with offices in USA and Sweden. Before that, I have worked as a research associate at Center of Artificial Intelligence and Computational Science in Information Technology University, Lahore. My research interests include Computer Vision, Signal Processing, and Computational Imaging.
I did my MS in Electrical Engineering from NUST Islamabad and BS from Institute of Space Technology, Islamabad.
Email  / 
Google Scholar  / 
LinkedIn
|
|
News
Two papers accepted in MICCAI.
One book chapter published in Deep Learning for Medical Image Analysis, The MICCAI Society Book Series.
Two papers accepted in CVPR.
One paper accepted in Medical Image Analysis.
One paper accepted in TPAMI.
One paper accepted in MICCAI.
Selected for Google Machine Learning Bootcamp for Train-The-Trainer Workshop.
One full-length paper related to Fourier Ptychography accepted in ICCV Workshop on Learning for Computational Imaging (LCI), 2019. The LCI workshop proceedings will be archived in the IEEE Xplore Digital Library and the CVF Open Access.
One paper accepted as poster in British Machine Vision Conference (BMVC), 2019.
Two posters accepted in CVPR-2019 workshop on Computational Cameras and Displays (CCD).
I am nominated as publicity ambassador of CVPR-19 workshop 'Computer Vision for Global Challenges' .
Paper accepted in Sampling Theory and Applications (SampTA), 2019.
Paper related to Fourier Ptychography accepted in ICASSP', 2019 (H-INDEX 71).
Paper submitted to BMVC, 2019.
Paper accepted in NIPS workshop of Machine Learning for Health, 2018.
Accepted in Udacity-Pytorch scholarship challenge, 2018
Secured NVIDIA GPU grant for Lab, 2017.
|
Research
I am interested in computational imaging, machine learning, optimization, and image processing.. Much of my recent research is about regularizing the inverse problems in signal and image processing using generative models (GANs, VAEs and Intertible models). Before that i have worked on theoretical aspect of samling architectures designed for low rank and sparse signals. I have also worked on removing Poisson noise in low light imaging in astronomy.
|
|
Blind Image Deconvolution using Pretrained Generative Priors
Muhammad Asim*, Fahad Shamshad* , Ali Ahmed
Submitted to 'Which-Must-Not-Be-Named', 2019
This paper proposes a novel approach to regularize the ill-posed and non-linear blind image deconvolution (blind deblurring) problem using deep generative networks
|
|
Blind Image Deconvolution using Deep Generative Priors
Muhammad Asim*,
Fahad Shamshad* ,
Ali Ahmed,
Submitted to TPAMI, 2018
arxiv /
blog post /
bibtex
In this paper, we employ generative models to regularize highly ill-posed blind image deblurring problem.
|
|
Deep Ptych: Subsampled Fourier Ptychography using Generative Priors
Fahad Shamshad,
Farwa Abbas,
Ali Ahmed,
Submitted to ICASSP, 2019
code /
bibtex
We propose robust ptychography algorithm that acheive comparable reconstruction results to state of the art at low subsampling ratios.
|
|
Poisson Denoising for Astronomical Images
Fahad Shamshad,
Mohsin Riaz,
Abdul Ghafoor,
Advances in Astronomy, 2018   (Spotlight)
supplement /
The scheme employs the concept of Exponential Principal Component Analysis and sparsity of image patches to remove high Poisson noise from astronomical images.
|
|
Robust Compressive Phase Retrieval using Generative Priors
Fahad Shamshad,
Ali Ahmed,
To be submitted to IEEE Transaction on Computational Imaging, 2018
project page
This paper proposes a new framework to regularize the highly ill-posed and non-linear phase retrieval problem through deep generative priors using simple gradient descent algorithm.p>
|
|