Islam Ebeid
Texas Woman's University, TX·Computer Science
Doctors rely on segmentation, the outlining of organs, tumors or other structures in medical scans, to make decisions and track disease. Deep-learning tools can automate this.
- Fields
Computer Science
Biomedical Engineering
- Topics
medical image segmentation
vision-language models
multimodal learning
- Grants
- 1 known grant, through Sep 2028
What they work on
- Develop an auto-prompting mechanism that uses graph-based reasoning to generate segmentation-specific text descriptions of medical images without human-written prompts.
- Build a dynamic knowledge-graph module, a network of linked medical facts, that encodes and updates expertise so segmentation generalizes across imaging modalities and anatomical regions.
- Design multi-level alignment of image and text features, with asynchronous fusion and bidirectional encoding, inside a closed loop where segmentation and annotation refine each other.
More about the research
But most of them look only at the image and miss the medical context a clinician brings. Vision-language models (VLMs), which link images with text, could add that context through written descriptions. However, they depend on hand-crafted prompts, are not well suited to segmentation, and generalize poorly across scan types and body regions.
Why it matters
More accurate and adaptable automatic segmentation could reduce the manual work of annotating medical images. It could also support clinicians' decisions and disease monitoring across different imaging types and body regions, with AI tools that hold up better in real clinical settings.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Yuxiong Wang
University of Illinois Urbana-Champaign, IL·Computer Science
Develop a multi-resolution video encoding strategy that adjusts token granularity to balance visual detail against efficiency in transformer-based vision-language models.
Jurij Leskovec
Stanford University, CA·Computer Science
Develop multi-level graph neural networks that handle graphs whose nodes are themselves complex structures, such as molecular graphs, for scientific foundation models.
Meng Jiang
University of Notre Dame, IN·Computer Science·Early-career
Develop and compare natural language generation models, one data-driven and one based on theories of abduction, for comparative summarization and argument generation in scientific domains.
The school
Texas Woman's University: 3 labs on PI Pages, $9M a year on research, 238 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.