Careers & Student Outcomes

Student Q&A: Ph.D Student Applies Deep Learning Algorithms to Biomedical Imaging

Xueshen Li ’25, who has authored multiple journal articles during his doctoral studies at Stevens, believes communication is key to actualizing his research ideas

In 2020, Stevens Institute of Technology Ph.D student Xueshen Li began his doctoral studies in the Intelligent Imaging and Image Processing Lab, directed by Yu Gan, assistant professor of biomedical engineering. Li, who relocated from the University of Alabama in 2021, continues his research in Gan’s lab today, implementing deep learning algorithms and language models to improve biomedical imaging through image reconstruction. 

Li joined Sophia Donskoy ’26, a second-year biomedical engineering student, to discuss the experience.

Students Xueshen Li and Xueshen Li stand next to each other in a laboratory - the Intelligent Imaging and Image Processing LabXueshen Li '25 and undergraduate biomedical engineering student-researcher Aaron Shamouil '25 in the Intelligent Imaging and Image Processing Lab

Donskoy: Can you describe your current doctoral research?

Li: My research interest is in deep learning and medical imaging, so in our lab, we are developing deep learning algorithms for an imaging modality known as optical coherence tomography (OCT). Currently, we have many exciting projects surrounding image reconstruction in biomedical imaging, as well as virtual staining. With virtual staining, we use OCT imaging to generate virtual histology images, which ultimately provides us with more information on these images. We are also working with large language models to assist clinical diagnosis using OCT imaging, which are currently state-of-the-art projects in medical imaging and deep learning. 

D: Can you describe how deep learning is implemented in your lab?

L: We use deep learning algorithms as customized tools, so to develop them, we first need to collect data as we train our algorithm. Based on different needs, we develop a particular algorithm to achieve each of the goals in the research, such as image reconstruction. From this, we are able to improve medical imaging by incorporating more information into the imaging data and enhancing image quality. 

D: What do you find most interesting in the work you are doing?

L: I find that my research is about always exploring. There are many existing technologies, yet so many of these can be improved. Being able to improve technology interests me because it means I am able to identify a problem and make improvements using my knowledge and skills set. For instance, two factors that can be improved in OCT imaging are increasing the range of information that can be acquired from the images, which we do with virtual staining by bringing color to gray-scale images, and improving the image quality, which is achieved through image reconstruction.

D: What do you find to be the most challenging?

L: There were several challenges I have encountered throughout doing research.The first was understanding new concepts. When I was new to the field, the ideas of deep learning were initially difficult to understand. This leads to the second challenge, which was the implementation of the concepts learned. Once I had a concept for my research, I had to develop an algorithm to implement it through programming and coding, which was not an easy task for me in the beginning. The third challenge was communicating with others. Before I joined the Ph.D program at Stevens, communicating with others was a struggle for me. However, I grew to learn that communication is necessary in order to articulate and push my ideas. Although this was the most challenging part for me in the beginning, improving my communication skills significantly helped me with furthering my research.

Xueshen Li stands at a podium in front of a large screen giving a presentation on his researchXueshen Li '25 presents his research at the Society of Photo-Optical Instrumentation Engineers (SPIE) conference.

D: What were some accomplishments in your research experience?

L: The most meaningful accomplishment for me was my first publication in 2021. I had no prior experience of writing a paper, so I had to work from scratch by collecting ideas, communicating with others, and ultimately implementing my knowledge and ideas. Now, our lab has several publications across a number of renowned journals, such as the Institute of Electrical and Electronics Engineers (IEEE) Transactions on Biomedical Engineering and Biomedical Optics Express. Currently, we have multiple ongoing projects that may lead to potential future publications, and our lab has also attended a number of research conferences, such as Photonics West. 

D: How has Stevens facilitated your work?

L: My experience at Stevens has certainly helped with developing my communication skills. Stevens has a very unique community, with students across many different backgrounds and a diverse faculty across many science and engineering fields. After I joined Stevens, I was involved in multiple projects, and within these projects, I have had to communicate and collect feedback from my collaborators. Ultimately, these collaborations uniquely shaped my experience at Stevens and have helped me grow as a researcher. 

D: What are your plans after graduation?

L: I plan to graduate within the next year, in 2025. After graduation, I hope to transition into industry, ideally pursuing fields pertaining to machine learning engineering or medical imaging analysis in hospitals. After completing multiple projects as a graduate student, I feel that I can uproot and further the practical applications of my research by pursuing career paths that explore areas related to my research field, such as language models applied in biomedical engineering.

Learn more about academic programs and research in the Department of Biomedical Engineering: