Ronald Stanley
Missouri University of Science and Technology, MO·Engineering
Melanoma, the most dangerous skin cancer, has risen in incidence for decades and is most curable when caught early. Dermoscopy, a magnified skin-surface imaging method, helps clinicians spot early melanoma, but specialist appointments take a long time and many front-line practitioners lack dermoscopy training.
- Fields
Engineering
Computer Science
Biomedical Engineering
- Topics
melanoma
skin cancer
dermoscopy
- Grants
- 1 known grant, through Mar 2027
What they work on
- Build an online platform with the International Skin Imaging Collaborative (ISIC) where experts worldwide annotate, segment and validate critical structures in dermoscopic skin lesion images.
- Identify which annotated lesion features contribute most to distinguishing melanoma from benign mimics, through feature analysis and validation studies on the curated image sets.
- Fuse whole-lesion deep learning with deep learning guided by localized expert annotations, and test whether the combination improves structure detection and diagnosis.
- Use statistical analysis to determine how large small-scale annotated feature sets must be, and compare detection and diagnostic rates across standard, guided and fused models.
More about the research
Computer programs that learn from images (deep learning) can sometimes match or exceed dermatologists, yet they are not yet accurate enough for reliable screening, and they struggle when only small sets of examples exist for specific lesion features.
Why it matters
More accurate automated reading of dermoscopy images could help catch melanoma at its most curable stage, especially for patients seen first by mid-level practitioners, in virtual consultations, or where specialist waits are long. A public, expert-annotated archive of lesion structures could give researchers a transparent foundation for building and comparing automated cancer image analysis and lesion triage tools.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Kuan Huang
Kean University, NJ·Computer Science
Develop weakly supervised breast cancer detection that locates lesions in ultrasound images without radiologist-drawn bounding boxes, using a transformer-based classification network that weighs relationships across the whole image.
Islam Ebeid
Texas Woman's University, TX·Computer Science
Develop an auto-prompting mechanism that uses graph-based reasoning to generate segmentation-specific text descriptions of medical images without human-written prompts.
Kazlin Mason
University of Virginia, VA·Medicine & Clinical Research
Develop a deep learning framework that automatically segments velopharyngeal structures in 3D and measures them in 2D from MRI scans of children with velopharyngeal anomalies.
The school
Missouri University of Science and Technology: 7 labs on PI Pages, $70M a year on research, 96 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.