Abdullah-Al-Zubaer Imran

Abdullah-Al-Zubaer Imran
Email: aimran [AT] Stanford [DOT] edu

I am a Postdoctoral Research Scholar in the Wang group in the Radiological Sciences Laboratory (RSL) at Stanford University.

I obtained my PhD in Computer Science from the University of California, Los Angeles (UCLA) in 2020 under the supervision of Distinguished Professor Demetri Terzopoulos. During my PhD, as a member of the UCLA Graphics and Vision Lab, I worked on developing effective deep learning techniques for medical image analysis.

Earlier, I received an MS in Computer Science degree from Delaware State University (DSU) in 2016 and a Bachelor in Computer Science and Engineering (CSE) degree from Rajshahi University of Engineering and Technology (RUET), Bangladesh in 2012.

Previously, I worked in the Medical Imaging and Simulation (MEDIS) Lab at DSU in collaboration with the Penn XPL Lab, jointly advised by Professor David Pokrajac and Professor Predrag Bakic. I have also spent time at Philips Corporate Research NA at Cambridge, MA and Tencent Medical AI Lab at Palo Alto, CA.

Check out:

News

Feb 2021: Received the Stanford Bio-X Travel award!

Feb 2021: Gave a guest lecture in the Advanced Data Mining class at Utah State University

Jan 2021: Two papers accepted at ISBI 2021!

Jan 2021: Presented our Progressive adversarial semantic segmentation paper at ICPR 2020

Dec 2020: Presented our Partly supervised multi-task learning paper at ICMLA 2020 [Slides]

Oct 2020: Gave a guest lecture in the Deep Learning class at New Mexico State University

Oct 2020: Got selected for the 2020 AAPM Expanding Horizons Travel Grant award!

Sep 2020: Gave an invited talk on Emerging Biomedical Imaging Technologies at RUET

Sep 2020: Full paper accepted at ICMLA 2020

Old News

Aug 2020: Got featured on DiscoverPhDs
Jul 2020: Presented our scoliosis analysis paper at CBMS 2020
Jul 2020: Joined Stanford as a Postdoc
Jun 2020: Paper accepted at ICPR 2020
Jun 2020: Gave an invited talk at AI4HC 2020
May 2020: Defended my PhD Dissertation!!
Apr 2020: Got selected to attend MLSS 2020, Tuebingen (13% acceptance)
Apr 2020: Full paper accepted at CBMS 2020
Apr 2020: Extension of our ICMLA 2019 paper accepted as a chapter for Deep Learning Applications, Volume 2 book
Feb 2020: Presented our work on self-supervised semi-supervised multitasking at AAAI 2020, New York
Jan 2020: Participated in the Deep Learning and Medical Applications Workshop at IPAM, UCLA
Jan 2020: Attended the PhD Summit at Google LA

Old Old News

Dec 2019: Presented our deep generative modeling paper at ICMLA 2019, Boca Raton, FL
Dec 2019: Presented our work on landmark detection and vertebrae segmentation at Med-NeurIPS 2019, Vancouver, BC
Dec 2019: Got awarded a scholarship for travel expenses to AAAI 2020
Oct 2019: Paper accepted to AAAI 2020 Student Abstract and Poster Program
Oct 2019: Presented our semi-supervised multitasking paper at MICCAI 2019 MLMI, Shenzhen, China
Oct 2019: Two extended abstracts accepted to Med-NeurIPS 2019
Sep 2019: Extension of our DLMIA 2018 paper accepted to the journal CMBBE: Imaging & Visualization
Sep 2019: Paper accepted to ICMLA 2019 for oral (28% acceptance)!
Aug 2019: Paper accepted to MLMI 2019
Jun 2019: Joined Tencent America-Medical AI Lab, Palo Alto (Research Internship)
Mar 2019: Advanced to PhD candidacy!
Mar 2019: Got accepted with financial support to the DLRL Summer School 2019, Alberta
Feb 2019: Received LabEx PRIMES fellowship to participate in DeepImaging 2019, Lyon
Sep 2018: Our paper on lung lobe segmentation won the NVIDIA Best Paper Award at DLMIA 2018
Sep 2018: Presented our pulmonary lobe segmentation paper at MICCAI 2018 DLMIA, Granada, Spain
Aug 2018: Attended the Machine Learning Summer School (MLSS) 2018, Madrid
Jul 2018: Paper accepted to MICCAI DLMIA 2018
Jun 2018: Joined Philips Research North America, Cambridge (Research Internship)

Research

My research is primarily centered around artificial intelligence, computer vision, and medical imaging. I am particularly interested in developing advanced AI-powered medical imaging tools for clinical applications. Recently, I focus on semi-supervised learning with multi-task learning, self-supervised learning, and generative modeling. I am also interested in visual representation learning and domain generalization for medical image analysis tasks.

Theses

From Fully-Supervised, Single-Task to Scarcely-Supervised, Multi-Task Deep Learning for Medical Image Analysis
PhD Thesis 2020 at the University of California, Los Angeles, California, USA
[Abstract] [Oral] [BibTex]

Estimation of Breast Anatomical Descriptors From Mastectomy CT Images
MS Thesis 2016 at Delaware State University, Dover, Delaware, USA
[Abstract] [Oral] [BibTex]

Automatic Extraction of Road Networks From High Resolution Satellite Images
BS Thesis 2012 at the Rajshahi University of Engineering & Technology, Rajshahi, Bangladesh
[Abstract] [Oral] [BibTex]

Selected Publications SSIQA: Multi-Task Learning for Non-Reference CT Image Quality Assessment With Self-Supervised Noise Level Prediction Abdullah-Al-Zubaer Imran, Debashish Pal, Bhavik Patel, Adam Wang IEEE International Symposium on Biomedical Imaging (ISBI) 2021 [Paper] [Oral] [BibTex] A self-supervised regularizer predicting noise index, jointly optimized for reference-free CT image quality assessment.
MultiMix: Sparingly Supervised, Extreme Multitask Learning From Medical Images Ayaan Haque, Abdullah-Al-Zubaer Imran, Adam Wang, Demetri Terzopoulos IEEE International Symposium on Biomedical Imaging (ISBI) 2021 [Preprint] [Poster] [BibTex] [Code] A novel saliency birdge module adapted semi-supervised method, exploiting consistency augmentation and multi-source data, for jointly learning diagnostic classification and anatomical structure segmentation.
Multi-Adversarial Variational Autoencoder Nets for Simultaneous Image Generation and Classification Abdullah-Al-Zubaer Imran, Demetri Terzopoulos Deep Learning Applications, Volume 2, 2021 [Chapter] [BibTex] Joint learning of image generation and semi-supervised classification using a novel deep generative model with multiple discriminators.
Partly Supervised Multi-task Learning Abdullah-Al-Zubaer Imran, Chao Hunag, Hui Tang, Wei Fan, Yuan Xiao, Dingjun Hao, Zhen Qian, Demetri Terzopoulos International Conference on Machine Learning and Applications (ICMLA) 2020 [Paper] [BibTex] [Oral] Self-supervised regularization for jointly learning medical image classification and segmentation in limited labeled data settings.
Progressive Adversarial Semantic Segmentation Abdullah-Al-Zubaer Imran, Demetri Terzopoulos International Conference on Pattern Recognition (ICPR) 2020 [Preprint] [BibTex] [Oral] [Poster] A domain generalization approach without requiring domain specific data for generalized and improved medical image segmentation.
Fully-Automated Analysis of Scoliosis from Spinal X-Ray Images Abdullah-Al-Zubaer Imran, Chao Hunag, Hui Tang, Wei Fan, Kenneth MC Cheung, Michael To, Zhen Qian, Demetri Terzopoulos IEEE 33rd International Symposium on Computer Based Medical Systems (CBMS) 2020 [Paper] [BibTex] [Oral] [Video] A fully-automated pipeline based on segmentation of scoliotic vertebrae to measure and classify severity of scoliosis from anterior-posterior spine X-rays.
Self-supervised Semi-Supervised Multi-Context Learning for the Combined Classification and Segmentation of Medical Images Abdullah-Al-Zubaer Imran, Chao Hunag, Hui Tang, Wei Fan, Yuan Xiao, Dingjun Hao, Zhen Qian, Demetri Terzopoulos AAAI Conference on Artificial Intelligence 2020 [Paper] [BibTex] [Poster] A novel semi-supervised multiple-task model leveraging self-supervision and adversarial training, applied to classification and segmentation of medical images.
Bipartite Distance for Shape-Aware Landmark Detection in Spinal X-Ray Images Abdullah-Al-Zubaer Imran, Chao Hunag, Hui Tang, Wei Fan, Kenneth MC Cheung, Michael To, Zhen Qian, Demetri Terzopoulos Medical Imaging Meets NeurIPS (Med-NeruIPS) 2019 [Paper] [BibTex] [Poster] To guide a CNN in the learning of spinal shape while detecting landmarks in X-ray images, a novel loss based on a bipartite distance (BPD) measure is proposed which consistently improves landmark detection performance.
Fast and Automatic Segmentation of Pulmonary Lobes From Chest CT Using a Progressive Dense V-Network Abdullah-Al-Zubaer Imran, Ali Hatamizadeh, Shilpa P. Ananth, Xiaowei Ding, Nima Tajbakhsh, Demetri Terzopoulos Computer Methods in Biomechanics & Biomedical Engineering: Imaging & Visualization 2019 [Paper] [BibTex] Reliable, fast, and fully automated lung lobe segmentation based on a Progressive Dense V-Network (PDV-Net). This method can segment lung lobes in one forward pass of the network, with an average runtime of 2 seconds using a single Nvidia Titan XP GPU.
Semi-Supervised Multi-Task Learning With Chest X-Ray Images Abdullah-Al-Zubaer Imran, Demetri Terzopoulos MICCAI Machine Learning in Medical Imaging (MLMI) 2019 [Paper] [BibTex] [Poster] [Code] A novel multi-task learning model for jointly learning a classifier and a segmentor, from chest X-ray images, through semi-supervised learning. In addition, a new loss function that combines absolute KL divergence with Tversky loss (KLTV) to yield faster convergence and better generarlization of the such models.
Automatic Segmentation of Pulmonary Lobes Using a Progressive Dense V-Network Abdullah-Al-Zubaer Imran, Ali Hatamizadeh, Shilpa P. Ananth, Xiaowei Ding, Nima Tajbakhsh, Demetri Terzopoulos MICCAI Deep Learning in Medical Image Analysis (DLMIA) 2018 [Paper] [BibTex] [Oral] Nvidia Best Paper Award First and fast such kind of model which achieved a Dice score of 0.939 on LIDC and 0.950 on LTRC datasets, significantly outperforming a 2D U-Net model and a 3D Dense V-Net.