Deep learning-based automated classification of endometrial lesions in IVF patients using hysteroscopic images
Marah Barakat, Rasha Massoud, Marwan Alhalabi · 2026
Read Deep learning-based automated classification of endometrial lesions in IVF patients using hysteroscopic images by Marah Barakat, Rasha Massoud, Marwan Alhalabi (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13482470.txt.
Description
Barakat, Massoud, and Alhalabi train deep learning models in 2026 to read hysteroscopy images for IVF patients, sorting views of the womb lining into normal tissue, polyps, and endometritis, an inflamed lining. Expert reading of these scope videos varies with experience, so the team built a decision support system from 1,500 annotated stills drawn from 200 clinical videos. Four established image networks were adapted and compared on the three-way task.
The models scored between 85 and 89 percent accuracy overall, and DenseNet-121 came out strongest across the detailed measures reported by the team. For polyps it reached 93 percent sensitivity, catching nearly all true cases, with a discrimination score of 98.8 percent on the standard curve in testing. It also hit 90 percent precision for endometritis, and heat-map overlays confirmed the networks stared at the tissue features clinicians use, though outside validation is still owed.
Catalog record
| CREATOR | Marah Barakat, Rasha Massoud, Marwan Alhalabi |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 42 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |