Hydronephrosis from ultrasound images, Hydronephrosis.
| Article DOI | http://dx.doi.org/10.1109/CRV.2018.00021 |
| Objective | To classify the severity of prenatal hydronephrosis from renal ultrasound images. |
| AI Approach | CNN |
| Data Source(s) | Single institutional series (773 patients, 2492 ultrasound images) |
| Model Input | 256 x 256 pixel images of renal ultrasound |
| Model Outcome | Classifying prenatal hydronephrosis: All grades High v. low risk hydronephrosis Also available vs. SFU grades and vs. controls |
| Model Metrics | All grades: accuracy: 51.5%; F1: 0.52 High vs. low risk hydronephrosis accuracy of 79.8%; F1: 0.80 |
| Model Usability | NA |
AI = Artificial intelligence, SFU = Society for Fetal Urology
AIPEDURO 
Leave a comment