Invited Speakers of MICAD2023
Prof. Qin Zhang
Tsinghua University, China
Qin Zhang graduated from Tsinghua University, Beijing, China, with BS., MS. and Ph.D. Degrees in nuclear engineering in 1982, 1984 and 1989 respectively. He was a visiting scholar with University of Tennessee, Knoxville, TN, USA, and University of California, Los Angeles, CA, USA, from 1987 to1989, working on system reliability engineering and intelligent fault diagnoses. He is now a professor of Institute of Nuclear and New Energy Technology and Department of Computer Science and Technology, Tsinghua University, emeritus member of China Association for Science and Technology, member of International Nuclear Energy Academy, fellow of China Association for Artificial Intelligence (CAAI) and director of the specialized committee for causality and uncertainty in AI of CAAI, consultant of the specialized committee for wise medical care of CAAI. He originally developed a new AI model called Dynamic Uncertain Causality Graph for fault and disease diagnoses.
Speech Title: A Transparent and Explainable AI to Aid Clinical Diagnosis in General Practice
Abstract: DUCG (Dynamic Uncertain Causality Graph) is a new AI model to graphically represent clinical knowledge with uncertainty and make probabilistic reasoning for clinical diagnosis in general practice. Compared with the data-driven AI models, DUCG is knowledge-driven, transparent and explainable. This presentation will show online how DUCG guides primary clinicians to make clinical diagnoses for 64 chief complaints covering more than 1000 diseases included in 48 diagnostic models, and how to collect clinical information or what medical checks should be done step by step for each individual patient. Meanwhile, how the clinical experts work with DUCG platform to construct the 48 diagnostic models will be presented. The 48 models are: cough and sputum, dyspnea, abdominal pain, diarrhea, hematemesis, nasal congestion, epistemic, hematochezia, nausea and vomiting, arthralgia, hemoptysis, fever, lower urinary tract symptoms (frequent urination, urgency, pain in urination, dysuria, polyuria, gross hematuria, leakage of urine), chest pain, jaundice, anemia, edema, obesity, wasting, sore throat, palpitation, fever in children, neck and low back pain (neck pain, low back pain, back pain) , dizziness, headache, constipation, rash, dysphagia, lymph node enlargement, cyanosis, numbness of extremities, vaginal bleeding, abnormal vaginal discharge, vulval pruritus, decreased menstruation or amenorrhea, abdominal distension, syncope, tinnitus, deafness, earache, acid reflux, heartburn, hiccup, belching, mass, oliguria or anuria, eye diseases (eye pain, redness, tears), visual acuity (abnormal vision, visual field defect, visual object deformation, color vision change, flash sensation, diplopia, strabismus). Except that the last two models are waiting for third-party verification, the diagnostic precisions verified by third-party hospitals for the other 46 models are all more than 95%, in which the precision for every disease (including uncommon diseases) is no less than 80%. So far, about one million diagnosis cases in real world have been performed, in which only 17 cases were incorrect and the mistakes in DUCG were found and corrected. No same incorrect diagnoses appear after the corrections. Statistics show that DUCG can increase the ability of primary clinicians to diagnose diseases several times more than without DUCG.
Asst. Prof. Weimin Zhou
Shanghai Jiao Tong University (SJTU), China
Weimin Zhou, Ph.D. is a Tenure-Track Assistant Professor at the Global Institute of Future Technology at Shanghai Jiao Tong University (SJTU). Before joining SJTU in 2022, he was a Postdoctoral Scholar in the Department of Psychological & Brain Sciences at the University of California, Santa Barbara (UCSB). Dr. Zhou received his Ph.D. in Electrical Engineering from Washington University in St. Louis (WashU) in 2020. During his Ph.D., he worked as a Research Assistant in the Department of Biomedical Engineering at WashU and a Visiting Scholar in the Department of Bioengineering at the University of Illinois Urbana-Champaign (UIUC). Dr. Zhou possesses broad expertise in imaging science, computational image formation, visual perception, and machine learning. Dr. Zhou is the recipient of the SPIE Community Champion Award and the SPIE Medical Imaging Cum Laude Award. He also serves as a peer reviewer for a variety of journals and a program committee member for SPIE Medical Imaging. Dr. Zhou has been active in publishing research articles in top-tier journals and conference proceedings, including IEEE Transactions on Medical Imaging, Medical Physics, Journal of Biomedical Optics, Journal of Medical Imaging, and SPIE Medical Imaging.
Speech Title: Machine Learning-Based Ideal Observer Computation for Task-Based Assessment of Medical Imaging Systems
Abstract: Task-based measures of image quality (IQ) have been widely employed to support the objective assessment of medical imaging systems by academia, industry, and FDA. Task-based measures of IQ quantify the ability of an observer to perform specific tasks in images. For signal detection tasks (e.g., tumor detection), the Bayesian Ideal Observer (IO) utilizes complete task-specific information in images and sets an upper limit of observer performance. The IO has been advocated for use in optimizing medical imaging systems because, in this way, the amount of task-specific information in the measurement data can be maximized. However, in most cases, the determination of the IO test statistic is analytically intractable. Recently, machine learning methods have been developed for approximating the IO for realistic image data. In this talk, I will present state-of-the-art machine learning methods for computing the IO for signal detection tasks. Supervised learning-based methods that employ artificial neural networks (ANNs) and a sampling-based method that employs Markov-Chain Monte Carlo with generative adversarial networks (MCMC-GAN) will be described. These methods represent powerful computational tools that permit the objective assessment and optimization of medical imaging systems.
Asst. Prof. Fei Gao
ShanghaiTech University, China
Dr. Fei Gao is an assistant professor in ShanghaiTech University, and PI of Hybrid Imaging System Laboratory (HISLab: www.hislab.cn). He received Chinese Government Award for Outstanding Self-Financed Students Abroad (2014), Springer Thesis Award (2016). He is currently serving as associate editors of several journals, including Photoacoustics, Medical Physics, Ultrasound in Medicine and Biology, IEEE Photonics Journal. He also serves as TPC member of IEEE Ultrasonics Symposium. He has published about 170 journal and conference papers with 2700+ citations. His interdisciplinary research topics include photoacoustic (PA) imaging physics (proposed non-line-of-sight PA imaging, passive PA effect, PA resonance imaging, phase-domain PA sensing, pulsed-CW hybrid nonlinear PA imaging, TRPA-TRUE focusing inside scattering medium, etc.), biomedical circuits and systems (proposed miniaturization methods of laser source and ultrasound sensors, delay-line based DAQ system, hardware acceleration for PA imaging, etc.), algorithm and AI (proposed frameworks such as Ki-GAN, AS-Net, Y-Net, EDA-Net, DR2U-Net, etc,), as well as close collaboration with doctors to address unmet clinical needs (Some prototypes are under clinical trials).
Speech Title: Intelligent Photoacoustic Imaging
Abstract: The intelligence of medical imaging equipment is an irreversible trend, which has greatly accelerated the diagnostic procedure using CT and MRI. Photoacoustic imaging (PAI), as an emerging biomedical imaging modality revealing molecular/functional information noninvasively in deep tissue, is expected to be empowered by advanced artificial intelligence techniques. This is named as intelligent PAI, i.e. iPAI. Basically, there are three important scientific problems to be solved: (1) how to achieve the intelligence of PAI hardware system; (2) how to achieve the intelligence of PAI image reconstruction and processing; (3) how the intelligence of PAI system add values to the clinical practice. In this talk, we will introduce (1) handheld and endoscopic photoacoustic probes that can achieve adaptive adjustable bright-field and dark-field illumination, as well as a ring-shaped photoacoustic tomographic imaging system with adjustable imaging field of view, which can achieve optimal design of light scheme and ultrasound detection sensitivity employed by intelligent algorithms. (2) The method of fusing analog circuits with deep learning image reconstruction to achieve single-pixel real-time photoacoustic imaging. (3) A new framework combining traditional photoacoustic image reconstruction algorithms with deep learning algorithms, improving accuracy while preserving generalization. (4) Hardware acceleration of photoacoustic image reconstruction algorithms and intelligent data acquisition system design based on FPGA platform. (5) Non-line-of-sight photoacoustic imaging to overcome skull’s aberration by detecting PA signals through temporal bones. In the last part, we will introduce several potential biomedical applications of iPAI, which can benefit from the intelligent system design and deep learning algorithms, followed by discussion of the challenges and future directions of iPAI.
Assoc. Prof. Yangang Wang
Southeast University, China
Yangang Wang is currently an associate professor in School of Automation at Southeast University (SEU). Before joining in SEU, he worked as a research scientist at Microsoft Research Asia (MSRA) from 2014 to 2017. He received his Ph.D. in 2014 from Department of Automation at Tsinghua University, advised by Prof. Qionghai Dai. His primary research area involves in Computer Vision, Computer Graphics and Virtual Reality. His most recent research interests are digital virtual human (DVH) and dynamic shape reconstruction with single or multiview cameras (a.k.a. markerless motion capture), including 3D/4D human capture, modeling and reconstruction, Hand pose acquisition, modeling and simulation, Body shape reconstruction, simulation and control, and Tiny motion or vibration sensing, measurement and analysis. He is a program committee member of AAAI2019-2022, Chinagraph2020-2022 and CICAI2022. He also serves as the paper reviewer of SIGGRAPH / TVCG / TPAMI / CVPR / ICCV / NeurIPS and etc.
Title: Vision-based 4D Cardiac Reconstruction
Abstract: 4D Cardiac Reconstruction aims to leverage medical image processing and computer vision technologies for the analysis of the heart's structure and function. The primary goal is to enhance the early diagnosis of cardiovascular diseases and formulate effective treatment strategies, thereby significantly impacting patient health management. However, the field faces challenges in data annotation, model representation, and information utilization. To overcome these hurdles, we commence the process with cardiac segmentation, incorporating view planning, progressively assembling static cardiac models, and finalizing dynamic cardiac reconstruction. This systematic approach aims to construct a more comprehensive and personalized digital cardiac model, offering robust support for individualized healthcare and precision treatment.
Prof. Huimao Zhang
The First Hospital of Jilin University, China
Prof. Huimao Zhang is the Chair of the Department of Radiology at the First Hospital of Jilin University, China. She also serves as vice chair of the Chinese Society of Radiology, standing member of the Chinese Radiologist Association, chair of the Jilin Medical Association in Radiology, chair of Data Science and Artificial Intelligence Committee of the 15th Session of the Chinese Society of Radiology and vice chair of Abdominal Committee of the 15th Session of the Chinese Society of Radiology. Her research interests are oncologic imaging with special interest for Lung cancer, colo-rectal cancer, liver cancer, AI in oncologic imaging and emergency imaging, focusing on radiological imaging database construction, multi-organ segmentation, big data mining and disease screening.
Title: Progress and outlook of AI in Medical Imaging
Abstract: Due to the aging population and the rising incidence of chronic diseases, China is facing a significant challenge of growing shortage of doctors. Over the past decade, artificial intelligence (AI) has revolutionized all aspects of healthcare, which has shown great potential in addressing the challenges posed by the growing shortage of radiologists. China's AI industry has entered a new stage of development. An overview of the significant progress of medical imaging Al in practical applications would be provided, including optimizing workflow and assisting in disease diagnosis and treatment. Subsequently, the problems we are facing and the future outlook would be discussed.
Assoc. Prof. Sang Hyun Park
Daegu Gyeongbuk Institute of Science and Technology, South Korea
Sang Hyun Park is an Associate Professor in the Department of Robotics and Mechatronics Engineering at the Daegu Gyeongbuk Institute of Science & Technology (DGIST), South Korea (https://www.dgist.ac.kr/en/). He obtained his Ph.D. in Electrical and Computer Engineering from Seoul National University in February 2014. Following that, he dedicated two years as a Postdoctoral Fellow in the Biomedical Research Imaging Center at the University of North Carolina (2014-2016) and another year as a Postdoctoral Fellow at SRI International (2016-2017). His primary research focuses on medical image analysis, computer vision, and machine learning, encompassing areas such as weakly supervised learning, few-shot learning, generative AI, and federated learning. Selected as a Fulbright visiting scholar in 2024, Sang Hyun Park actively contributes to the research community. He served as an area chair of MICCAI 2023 and has been a co-organizer of the Predictive Intelligence in Medicine (PRIME) workshop at MICCAI for the past six years. He is also participating as an organizing team for MICCAI 2025 held in Korea. Moreover, he has undertaken the role of a reviewer for esteemed conferences and journals, including MICCAI, AAAI, CVPR, ECCV, TMI, and MedIA, over the last three years.
Title: Artificial Intelligence for Multi-Institutional Data Learning
Abstract: Artificial Intelligence (AI) is revolutionizing the field of medicine by providing advanced tools and techniques for medical image analysis. Recently, foundation models trained on large-scale datasets have shown high performance; however, it's challenging to train such models in the medical domain. Medical data is decentralized across various institutions, complicating the creation of substantial training datasets, and labels are frequently incomplete. Moreover, images obtained from diverse domains may manifest variations in characteristics. In this talk, I will introduce few-shot learning, domain adaptation, and federated learning techniques as innovative solutions to surmount these challenges in the field of medical image analysis.
Assist. Prof. Sujatha Krishnamoorthy
Wenzhou-Kean University, China
Sujatha krishnamoorthy is the Assistant professor of Computer Science at Wenzhou -kean University . Earlier to WKU she was the Research and Development Head at Sri krishna college of Engineering and Technology , Coimbatore ,India. she has over 7 years of research experience and over a decade long experience in teaching . Her specialization is Digital image processing with Image fusion. She has published over 60 papers in International refereed journals like Springer and Elsevier. She has delivered several guest lectures, seminars and chaired a session for various Conferences. She is serving as a Reviewer and Editorial Board Member of many reputed Journals and acted as Session chair and Technical Program Committee member of National conferences and International Conferences .She has received a Best Researcher Award during her research period.
Title: Detection and Classification of Diabetic Retinopathy(DR) using Machine Learning Techniques
Abstract: Diabetic Retinopathy (DR) primarily affects a set of lesions in the eyes, causing retinal degeneration and loss of vision. The DR features serve as a crucial component for ophthalmologists to diagnose DR at an earlier stage. This paper presents an automatic DR screening tool using a hybrid GAN-RCNN architecture formulated to categorize and identify different DR grades from the fundus images captured at various resolutions. The hybrid GAN-RCNN architecture is formulated by replacing the discriminator in the GAN with the RCNN classifier. The RCNN model can handle the complex inter-class and intra-class variations present in the fundus retina images and classify them into different classes such as mild, moderate, severe, and nonproliferative DR. The RCNN model not only extracts the pixels present in the fundus image but also focuses on the significant relationship that exists between different DR classes. The Archimedes optimization Algorithm (AOA) is used to optimize the different GAN and RCNN hyperparameters. When compared to the existing techniques the proposed model offers an accuracy of 99%, 98.5%, and 99.4% in the APTOS, Kaggle, and Messidor datasets which is comparatively high. The experimental outcome reveals that the introduced model serves as a concrete baseline for the diagnosis and treatment of DR patients.