Mohammad Al Fahim

OCUCFormer: An Over-Complete Under-Complete Transformer Network for Accelerated MRI Reconstruction

Mohammad Al Fahim, Sriprabha Ramanarayanan, Rahul G. S., Matcha Naga Gayathri, Arunima Sarkar, Keerthi Ram, Mohanasankar Sivaprakasam

Submitted to Computers in Biology and Medicine, 2023

Fully Automated Sinogram-based Deep Learning Model for Detection and Classification of Intracranial Hemorrhage

Chitimireddy Sindhura, Mohammad Al Fahim, Phaneendra K. Yalavarthy, Subrahmanyam Gorthi

Medical Physics, 2023

[paper]

HyperCoil-Recon: A Hypernetwork-based Adaptive Coil Configuration Task Switching Network for MRI Reconstruction

Sriprabha Ramanarayanan, Mohammad Al Fahim, Rahul G. S., Amrit Kumar Jethi, Keerthi Ram, Mohanasankar Sivaprakasam

IEEE/CVF International Conference on Computer Vision (ICCV) Workshop on Computer Vision for Automated Medical Diagnosis (CVAMD), 2023

[paper], [code]

Geometric Learning-Based Transformer Network for Estimation of Segmentation Errors

Sneha Sree Cumsali, Mohammad Al Fahim, Keerthi Ram, Mohanasankar Sivaprakasam

International Conference on Medical Image Computing & Computer Assisted Intervention (MICCAI) Workshop on Shape in Medical Imaging (ShapeMI), 2023

[paper]

SDLFormer for Accelerated MRI Image Reconstruction

Rahul G. S., Sriprabha Ramanarayanan, Mohammad Al Fahim, Keerthi Ram, Mohanasankar Sivaprakasam

International Conference on Medical Image Computing & Computer Assisted Intervention (MICCAI) Workshop on Medical Image Learning with Limited & Noisy Data (MILLanD), 2023

[paper], [code]

SFT-KD-Recon: Learning a Student-friendly Teacher for Knowledge Distillation in Magnetic Resonance Image Reconstruction

Matcha Naga Gayathri, Sriprabha Ramanarayanan, Mohammad Al Fahim, Rahul G. S., Keerthi Ram, Mohanasankar Sivaprakasam

International Conference on Medical Imaging with Deep Learning (MIDL), 2023

[paper], [code]

Detection, Classification and Segmentation of Traumatic Brain Injuries

Implemented a state-of-the-art deep learning model in PyTorch for detection and classification of Intracranial Hemorrhages (ICH) in CT scans

UNet++ architecture-based model was developed in Keras and Tensorflow for 2D and 3D semantic segmentation of ICH regions

Telugu Handwritten Text Recognition

Developed a Convolutional Recurrent Neural Network model to recognize Telugu text from handwritten text images.

Used conventional image processing techniques to capture bounding boxes around the words, and individual word images were fed to the model for recognition. The model was developed with Keras and Tensorflow in Python

StarGAN-v2 for Synthesis of DCE Prostate

Implemented StarGAN-v2 to synthesize Dynamic Contrast Enhanced Prostate images given T2, Proton-Density, and Diffusion-weighted MRI protocol images.

StarGAN-v2 was trained in PyTorch to generate Prostate MRI images of half-diffusion and full-diffusion of contrast agents in the Prostate, highlighting the cancerous regions.

Classification and Segmentation of White Blood Cells

Implemented several state-of-the-art deep learning architectures for classifying White Blood Cells (WBC)

To segment WBCs, we implemented unsupervised and weakly-supervised methods like K-Means clustering and Graph-Cuts methods due to the unavailability of segmentation maps. The models were implemented with Keras and Tensorflow in Python.