my photo

Saeed Seyyedi
Ph.D., M.Sc.

Computational and Data Science Specialist | AI and Deep Learning Researcher | J&J Fellow
Berkeley Institute for Data Science, University of California, Berkeley
Bakar Computational Health Science Institute , University of California, San Francisco
Curriculum Vitae
Google Scholar
sd.seyyedi [at]
490 Illinois Street
San Francisco, CA 94158



AACR Virtual Special Conference on Artificial Intelligence, Diagnosis, and Imaging, Jan 13-14, 2021


SPIE Medical Imaging'20
SPIE Medical Imaging Conference, Houston, TX, Feb 15-20, 2020

Virtual Annual Meeting, Society for Imaging Informatics in Medicine, Jun 24-26, 2020

Conference on Machine Intelligence in Medical Imaging, Sep 13-14, 2020


Conference on Machine Intelligence in Medical Imaging, Austin, TX, Sep 22-23, 2019

World Conference on Lung Cancer, Barcelona, Spain, Sep 7-10, 2019


AACR Artificial Intelligence Conference, Newport, RI

BC Cancer Summit'18
British Columbia Cancer Summit, Vancouver, Canada

World Conference on Lung Cancer, Toronto, Canada


International Fully3D Conference, Xi'an, China


International Symposium on Biomedical Imaging, Prague, Czech Republic

CT Meeting'16
Bamberg, Germany


Advanced Imaging and Visualization, Austria


International Symposium on Biomedical Imaging, Chicago, IL


E-Health and Bioengineering Conference, Iasi, Romania

Symposium on Signal and Image Processing, Trieste, Italy

Flag Counter

Dr. Saeed Seyyedi is a Computational and Data Science Research Specialist and a Johnson & Johnson Fellow in the Berkeley Institute for Data Science at University of California, Berkeley, and Bakar Computational Health Science Institute at Unersity of California, San Francisco. Dr. Seyyedi is a computer scientist and biomedical engineer and has specialized in development and deployment of artificial intelligence models including Deep Learning and Machine Learning for multi-modality real world vision, imaging and unstructured data. Dr. Seyyedi received his Master of Science in Biomedical Engineering where he studied techniques for processing and analysis of digital breast tomosynthesis data. He obtained his PhD in Computer Science, Medical Imaging and Informatics from Technical University of Munich in Germany where he was the recipient of an E.U. Marie Curie research fellowship supporting his research and studies in the field of computer science, medical imaging and informatics. During his PhD studies, he was a visiting scholar at Johns Hopkins University where he was involved in development of advanced models for multi-modality imaging problems. Prior to joining UCSF and UCB, he has worked in several academic and industry centers. As a postdoctoral research fellow at Stanford University, he led several projects in AI and medical imaging research, with a particular interest in development of deep learning and computer vision methods to detect and characterize cancer on radiologic images. Additionally, he served as an imaging data scientist at AstraZeneca, where he led multiple projects and collaborations with several academia and industry groups focusing on AI based applications for analysis of big medical and biological datasets including digital pathologic and radiologic imaging data. He also worked with the British Columbia Cancer Agency of Canada as a part of PanCan lung cancer screening project where he developed deep learning and radiomics applications for lung cancer detection and classification. He is an author, reviewer and editorial board member in several journals and conferences and his interdisciplinary research and studies have been supported and recognized by a number of awards and grants.

News and Updates

10/2020 Our Nature Digital Medicine paper on Multi-modality Deep Learning for Medical imaging and EHR is now available online
10/2020 I am serving as a reviewer for 2020 Conference on Neural Information Processing Systems (NeurIPS 2020)
09/2020 Our paper "SCREENet: A Multi-view Deep Convolutional Neural Network for Classification of High-resolution Synthetic Mammographic Screening Scans" is available on Arxiv!
09/2020 Our abstract on building an AI model to improve berast cancer screening was presented in the Conference on Machine Intelligence in Medical Imaging
08/2020 I am serving as a program committee member at the 4th International Conference on Bioscience & Engineering (BIEN 2020)
07/2020 I am serving as a program committee member at the 17th ACS/IEEE International Conference on Computer Systems and Applications (AICCSA 2020)
05/2020 I will be serving as a technical committee member at the 4th International Conference on Biomedical Engineering and Bioinformatics (ICBEB 2020)
03/2020 I will be serving as a technical committee member at the 2nd International Conference on Intelligent Medicine and Health (ICIMH 2020)
03/2020 I am serving as a reviewer for 23rd International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2020)
01/2020 I am serving as a reviewer for Journal of Computerized Medical Imaging and Graphics
10/2019 Our paper is accepted at 2020 SPIE Medical Imaging Conference at Houston, TX
09/2019 I am serving as a reviewer for 2019 Conference on Neural Information Processing Systems (NeurIPS 2019)
08/2019 I am serving as a reviewer for IEEE Journal of Transactions on Biomedical Engineering
07/2019 I joined as an editorial board member for Journal of Medical Artificial Intelligence (JMAI)
07/2019 I am serving as a reviewer for Journal of BMC Cancer Imaging
06/2019 My proposal is shortlisted as a finalist for Stanford School of Medicine Dean's Postdoctoral Fellowship
06/2019 I am serving as a reviewer for 2019 AIMI seed grant awards at Stanford Unievrsity
05/2019 our paper is accepted at 2019 World Conference on Lung Cancer at Barcelona, Spain

I am actively looking for students who are interested in getting involved in AI research in medical imaging. Please email me if you are familiar with machine learning and deep learning methods and are interested to apply those methods to tackle challenges in real world imaging problems.


Lung Cancer Detection and Classification using Deep CNNs and Transfer Learning

With the recent recognition that lung cancer screening using LDCT can improve the mortality of this disease new clinical challenges have arisen. The classification of sub-cm lung nodules and prediction of their behavior presents a challenge for physicians and computer aided diagnosis. In this research, we developed several deep convolutional neural networks (CNNs), transfer learning and radiomics based machine learning techniques to aid in the detection, classification and management of small lung nodules.

publications: Comparison of Classical Machine Learning and Convolution Neural Nets for the Differentiation of Malignant from Benign Sub 1.1 mm Lung Nodules in CT Scans @AACR
Machine Learning and Deep Learning Approaches for Classification of Sub-cm Lung Nodules in CT Scans @SPIE
Optimizing Radiomics Features by Minimizing Boundary Effects and Normalizing with Opposite Lung Tissue Characteristics @JTO


Digital Pathology Image Understanding using Deep Learning Models

In this project, we design and develop deep learning and multiple instance learning based models and image understanding methods to help pathologists in diagnosis and management of patients based on pathology data and images.


Low-dose CT Perfusion of the Liver using RoD

CT perfusion imaging of the liver enables the evaluation of perfusion metrics that can reveal hepatic diseases and that can be used to assess treatment responses. However, x-ray radiation dose limits more widespread adoption of liver CT perfusion studies as a diagnostic tool. In this study, we developed a model-based reconstruction method called Reconstruction of Difference (RoD) for use in low-dose CT perfusion of the liver which integrates a baseline non-contrast-enhanced scan into the reconstruction objective to improve image volumes formed from low-exposure data. Specific perfusion metrics include hepatic arterial perfusion (HAP), hepatic portal perfusion (HPP) and perfusion index (PI) parameters computed using the dual-input maximum slope method (SM). The quantitative and qualitative comparisons of reconstructed images and perfusion maps shows that the RoD approach can significantly reduce noise in low-dose acquisitions while maintaining accurate hepatic perfusion maps as compared with traditional reconstruction methods.

publications: Low-Dose CT Perfusion for the Liver using Reconstruction of Difference @IEEE-TRPMS
@Fully3D 2017


X-ray Tensor Tomography: Acquisition, Reconstruction and Noise Reduction

In this project, we present X-ray tensor tomography (XTT) imaging modality using a compact laboratory setup that reveals information about micrometer-sized structures within samples that are several orders of magnitudes larger. XTT is a novel imaging modality for the three-dimensional reconstruction of X-ray scattering tensors from dark-field images obtained in a grating interferometry setup. The two-dimensional dark-field images measured in XTT are degraded by noise effects, such as detector readout noise and insufficient photon statistics, and consequently, the three-dimensional volumes reconstructed from this data exhibit noise artifacts. In the second part of this project, we developed iterative reconstruction schemes with incorporated denoising steps i.e., the popular total variation denoising technique. For the forst time, we used XTT to measure femur bone sample and carbon fibre sample and applied the proposed methods to acquire images, reconstruct and denoising the data.

publications: Six dimensional X-ray Tensor Tomography @Applied Physics Letters
Reconstruction and Noise Reduction @IEEE-TCI
Component-based Reconstruction Method @ISBI 2016


Digital Breast Tomosynthesis: Acquistion and Image Reconstruction

Digital breast tomosynthesis (DBT) is an innovative imaging modality that provides 3D reconstructed images of breast to detect the breast cancer. In this project, we developed multiple tomograpic reconstruction techniques to improve the quality of resultant three-dimensional images. We also studied the effect of different acquisition schemes on image quality when applied with iterative reconstruction techniques

publications: Object-Oriented Simulator for 3D DBT Imaging System @Computational and Mathematical Methods in Medicine
image reconstruction using anisotropic total variation minimization @IEEE
Evaluating the effect of acquisition parameters in DBT @IEEE
3-D Tomosynthesis Image Reconstruction Using Total Variation @BioMedCom