Overview
Join us for the next NISS-CANSSI Collaborative Data Science Webinar on May 8, 2025, from 1 to 2 PM ET, titled "From MPEG-4 to Deep Learning: Transforming Audio-Visual Analytics for Healthcare and Beyond." This webinar will explore the evolution of audio-visual analytics from traditional compression standards like MPEG-4 and H.264 to modern deep learning approaches that enable real-time object detection, action recognition, and intelligent healthcare applications. Advancements in deep learning have significantly redefined how we process image, video, and audio data, making possible innovations such as robotic patient monitoring and voice-activated medical assistants. Often referred to as "orange technology," these integrated techniques are enhancing the quality of life in medical settings and transforming the landscape of healthcare and beyond. Don’t miss this opportunity to gain insights into how classical and modern methods converge to shape the future of audio-visual data science.
Speakers
An-Chao Tsai, Ph.D., SMIEEE Associate Professor, National Pingtung University, Taiwan
Anand Paul, LSU Health-New Orleans
Moderator
Qingzhao Yu, Associate Dean for Research at the School of Public Health, Louisiana State University Health, New Orleans
Abstract
Title: Title : From MPEG-4 to Deep Learning: Transforming Audio-Visual Analytics for Healthcare and Beyond
Over the past two decades, image and video processing has undergone a revolutionary transformation. Traditional standards such as MPEG-4 and H.264 laid the groundwork for efficient storage and transmission by exploiting motion estimation, block-based transforms, and predictive coding. More recently, advanced deep learning models have significantly redefined the field, offering higher accuracy and deeper insights into visual data. For example, architectures like YOLO enable real-time object detection and facilitate tasks like region-of-interest tracking, action recognition, and video summarization. Simultaneously, multi-view coding techniques—once reliant on handcrafted features—now benefit from data-driven optimization that leverages spatiotemporal coherence. This has led to robust motion estimation, improved inter/intra-frame predictions, and an expanded range of applications well beyond compression. Additionally, integrating deep learning–based image and video processing with sound signal analysis paves the way for robotic agents in healthcare, using camera feeds for patient monitoring and acoustic signals for voice-activated commands. Often termed “orange technology,” these synergistic approaches enrich quality of life in medical settings. This webinar explores the shift from classical compression to advanced deep learning and highlights how these paradigms converge to drive innovations in healthcare, robotics, and beyond.
About the Speakers
Dr. An-Chao Tsai received his Ph.D. in Electrical Engineering from National Cheng Kung University, Taiwan, in 2010. He is currently an Associate Professor in the International Master Program of Information Technology and Application at National Pingtung University, Taiwan. His research focuses on artificial intelligence, virtual reality, and AIoT. Dr. Tsai has contributed extensively to AI-driven medical applications, agricultural precision analysis, and computer vision. His recent work includes AI-based skin analysis using conditional generative adversarial networks for melasma diagnosis, real-time classification of black soldier fly larvae for sustainable food waste management, and intelligent IoT-based farming systems for optimizing agricultural productivity. As a Senior Member of IEEE, Dr. Tsai has served as the Track Chair and Program Chair for the IEEE International Conference on Orange Technologies since 2015. His interdisciplinary research integrates AI, IoT, and deep learning for practical, real-world applications in healthcare, smart agriculture, and education.
Dr. Paul Anand is an Associate Professor in the Department of Biostatistics and Data Science at the School of Public Health, Louisiana State University Health Sciences Center. He earned his Ph.D. in Electrical and Computer Engineering from National Cheng Kung University, Taiwan, R.O.C., in 2010. His research interests encompass Big Data Analytics, Artificial Intelligence (AI), and Machine Learning, with a particular focus on Resilient and Robust Intelligence, including Generative AI and Artificial General Intelligence. Dr. Paul was recognized among the top 2% of scientists worldwide by Stanford University and Elsevier Publisher for the years 2022 and 2024. He has been an IEEE Senior Member since 2015. In addition to his research, Dr. Paul has served as an editor for several prestigious SCIE journals, including IEEE Access, Computer Animation and Virtual Worlds, ICT Express, PeerJ Computer Science, Cyber-Physical Systems (Taylor & Francis), International Journal of Interactive Multimedia and Artificial Intelligence, and ACM Applied Computing Review. He has also held the role of Track Chair for Smart Human-Computer Interaction at ACM SAC from 2014 to 2019.
About the Moderator
Dr. Qingzhao Yu is a biostatistics and data science professor and the associate dean of research at the School of Public Health, LSU Health-New Orleans. As a researcher, Dr. Yu has developed statistical methods in causal inference, clinical trials, Bayesian methods, spatial analysis, data mining, and machine learning methods. Her research areas of interest include health disparities, cancer, and chronic diseases. Dr. Yu has published over 140 methodology and collaboration papers and five software packages. She is a co-editor for the journal Data Science in Science. Dr. Yu is a PI and Co-Investigator for multiple grants supported by NIH, CDC, and other national and state funding agencies.
About the NISS-CANSSI
Collaborative Data Science Web Series:
The NISS-CANSSI Collaborative Data Science initiative that the National Institute of Statistical Sciences (NISS) in collaboration with the Canadian Statistical Sciences Institute (CANSSI) brings together experts from various fields to tackle complex data challenges through interdisciplinary teamwork and innovative methodologies.
Goals of the Initiative
The goal is to foster progress in:
- Developing new ideas for experimental and observational data-driven learning and discovery that address key questions at the cutting edge of science and scientific deduction;
- Quantifying and summarizing uncertainty in data-driven theories, as well as complex Data Science models, algorithms, and workflows; and
- Establishing new practices for scientific reproducibility and replicability through Data Science.
Featured Webinars
Changing Climate, Changing Data: A journey of statisticians and climate scientists
Date: Thursday, March 20, 2025 at 1-2pm ET
Speakers: Claudie Beaulieu, Assistant Professor of Ocean Sciences, University of California, Santa Cruz and Rebecca Killick, Professor of Statistics, School of Mathematical Sciences, Lancaster University; Moderator: Emily Casleton, Statistical Sciences Group, Los Alamos National Laboratory (LANL)
Astronomy & Cosmic Emulation
From MPEG-4 to Deep Learning: Transforming Audio-Visual Analytics for Healthcare and Beyond
Data Science Techniques for Control of Assistive Devices After Neurological Injury
Event Type
- NISS Hosted
Host
Sponsor
Website
Location
Policy
