Dr. Youshan Zhang, assistant professor of artificial intelligence and computer science in the Katz School of Science and Health;
Dave DeFusco
Dr. Youshan Zhang, assistant professor of artificial intelligence and computer science at the Katz School of Science and Health, is an innovative AI-powered diagnostic tool.
Cardiomegaly, the enlargement of the heart, is an important early indicator of heart disease, especially in dogs, and is one of the leading causes of death in humans and animals. Traditionally, detection of this condition has relied on manual analysis of chest radiographs, also known as chest radiographs, using the vertebral heart scale (VHS). However, this method is time-consuming, prone to human error, and requires specialized knowledge, making it inefficient and inaccessible for widespread use.
Dr. Zhang’s project, “Heart Disease Detection with AI for Veterinary Medicine,” addresses these challenges by developing deep learning models that can automate the VHS process more accurately and faster.
“The main goal of this project is to bridge the gap between traditional clinical methods and advanced AI models,” said Dr. Zhang. “Many clinicians, especially those without deep learning experience, struggle to trust AI-generated results because current models lack transparency and explainability.”
To build reliability and ease of use, this project aims to integrate traditional VHS metrics into a deep learning framework. This will enable clinicians to better understand how AI-derived predictions align with established medical standards. Research will focus on improving the transparency and accuracy of these models, making the diagnostic process more intuitive for veterinarians and reducing reliance on manual calculations.
This project builds on Dr. Zhang’s previous work published in Scientific Reports, in which he introduced a regressive visual transformer (RVT) for cardiac hypertrophy assessment in dogs. Building on this foundation, his new project outlines three main goals.
Developing a new detection model: Dr. Zhang will create a new tool called the Vertical Fully Connected Layer (PFCL). This ensures that the way the heart is measured in X-ray images is more accurate by ensuring that specific lines are used. The measurements are perfectly perpendicular to each other. This allows computer models to better detect important features of the heart and calculate its size more accurately, increasing confidence in diagnosing heart problems.
Automatic report generation: Using deep semantic mapping and few-shot generation techniques, Dr. Zhang will develop a tool that can generate cardiac hypertrophy reports with minimal training data. This streamlines the diagnostic process, especially the initial assessment by general practitioners.
User-friendly software interface: The main outcome of this project is the creation of a software interface that is easy to access for clinicians and the general public. The tool integrates data labeling, outcome prediction, report generation, and modification into one platform, making it easy to use without prior domain knowledge.
“By developing more accurate and accessible diagnostic tools, this project aims to reduce the cost of detecting cardiac hypertrophy, while improving diagnostic accuracy and reducing stress for pet owners.” Dr. Zhang said. “The project’s deep learning models could also pave the way for similar AI applications in human medicine, particularly in improving early detection of heart disease.”
