Haowen Xue

Title: Haowen Xue - Innovator in Retinal Blood Vessel Segmentation

Introduction

Haowen Xue is a distinguished inventor based in Jiangsu, China, who has made significant contributions to the field of medical imaging, specifically in the segmentation of retinal blood vessels. With a focus on enhancing the accuracy of medical diagnostics, his work addresses critical challenges in retinal imaging.

Latest Patents

Haowen Xue holds a patent for the "Three-way U-Net Method for Accurately Segmenting Uncertain Boundary of Retinal Blood Vessel." This innovative method provides a solution for accurately identifying the uncertain boundaries of retinal blood vessels. It involves outlining the uncertainty in blood vessel boundary labels, constructing upper and lower bounds using dilation and erosion operators, and devising a three-way loss function to train network parameters using stochastic gradient descent. The patent also outlines the development of an auxiliary diagnosis application system capable of intelligently segmenting retinal blood vessels, incorporating fundus data acquisition and diagnostic support features.

Career Highlights

Currently, Haowen Xue is associated with Nantong University, where he continues to enhance his research and development in the field of ophthalmic imaging. His academic and professional endeavors have focused on the intersection of technology and medicine, facilitating advancements in patient diagnosis.

Collaborations

In his research, Haowen has collaborated with esteemed colleagues Weiping Ding and Ying Sun. Together, they work on innovative projects that aim to push the boundaries of medical imaging and improve diagnostic accuracy in healthcare.

Conclusion

Haowen Xue’s contributions through his innovative patent and collaborative efforts at Nantong University underscore his dedication to advancing medical technologies. His work in retinal blood vessel segmentation not only enhances diagnostic capabilities but also embodies the spirit of innovation within the field of medical imaging.