Embedded Files
Dra. Graciela María de Jesús Ramírez Alonso
Graciela María de Jesús Ramírez Alonso is a full-time professor in the Faculty of Engineering at the Universidad Autónoma de Chihuahua, where she leads research projects focused on developing deep learning models for image and signal analysis, with an emphasis on medical applications. She holds a degree in Electronic Engineering and a Ph.D. in Electronic Engineering and is a Level I member of Mexico’s National System of Researchers (SNII).
She has established a productive research program, with publications in journals indexed in the Journal Citation Reports (JCR). Her work integrates artificial intelligence techniques for image classification, segmentation, and detection, automatic report generation from medical images, and the analysis of EEG and biometric signals.
She actively supports the development of future researchers by supervising undergraduate and graduate theses. She considers participation in national and international conferences essential to sharing research, building collaborative networks, and enriching academic practice.
She currently coordinates the Computer Vision and Data Science Laboratory, an interdisciplinary space that advances research with a social focus, aiming to improve AI-assisted medical diagnosis, particularly in clinical settings with limited resources.
Title: Understanding the Data, Interpreting the Metrics: Medical Report Generation from Images
Abstract:
This talk explores the generation of medical reports from medical images, emphasizing the importance of understanding the data used to train and evaluate models. Drawing on our work in the ImageCLEF 2026 challenge, it highlights the role of data preparation and text preprocessing, which often receive less attention than model design. It examines how inconsistencies and limitations in image–text datasets can place additional demands on models and complicate the interpretation of their results. The talk also raises questions about the relationship between evaluation metrics and the quality of generated medical text, including whether higher scores necessarily indicate better outputs. Throughout the talk, our implemented approaches and experimental findings illustrate lessons learned from data preparation and model development.
Dr. Eduardo Galicia Gómez
Dr. Eduardo Galicia is an AI Scientist and Developer with a PhD in Computer Science and Engineering from IIMAS-UNAM. Specializing in Medical AI, his core expertise lies in ultrasound and MRI image processing. Currently at AISHA, he focuses on designing and training deep learning models for full-body segmentation. By combining this domain knowledge with a robust tech stack—featuring Python, cloud infrastructure, containerization, and 3D medical tools—he builds scalable, end-to-end AI solutions that bridge the gap between AI research and clinical deployment.
Title: Scaling Full-Body MRI Segmentation: A Cloud Architecture Approach
Dr. Shumin Li
Associate Professor, Jilin University; Head of Veterinary Solutions, Xiaowen Tech/VetMind
Shumin Li is a veterinarian and data scientist specializing in the integration of artificial intelligence within veterinary medicine. She holds a Doctor of Veterinary Medicine from Kansas State University and a Master of Computer and Information Technology from the University of Pennsylvania. With extensive clinical and research experience across the United States, Hong Kong, France, and Taiwan, Dr. Li has worked in diverse environments ranging from specialty emergency centers to federal veterinary management. She currently focuses on predictive diagnostics and analytics in clinical veterinary medicine, positioning her as a rare expert capable of bridging the gap between deep clinical veterinary practice and advanced AI innovation.
Title: Building Evidence-Grounded Multimodal AI for Veterinary Clinical Decision Support
Abstract:
Artificial intelligence is rapidly transforming medical imaging and clinical decision-making. However, moving from isolated AI predictions to reliable clinical decision support requires more than accurate image recognition. It demands the integration of multimodal patient data, domain-specific medical knowledge, and evidence-grounded clinical reasoning.
In this talk, Dr. Shumin Li will explore the development of veterinary-specific AI systems, using VetMind's clinical decision support (CDS) framework as a practical example. The presentation will examine how large language models can integrate unstructured clinical records, laboratory findings, diagnostic imaging, and structured medical knowledge to support differential diagnosis, diagnostic planning, and treatment recommendations. Particular attention will be given to knowledge engineering, hybrid retrieval architectures combining vector databases and knowledge graphs, and training strategies designed to align AI-generated recommendations with established clinical evidence.
The talk will also address critical challenges in translating multimodal AI into clinical practice, including explainability, hallucination mitigation, clinical validation, and workflow integration. Drawing on experience at the intersection of veterinary medicine and AI development, Dr. Li will highlight how evidence-grounded multimodal systems may evolve from standalone predictive tools into clinically meaningful decision-support partners.
Page updated
Google Sites
Report abuse