Publications by authors named "Lia Cmj Goltstein"
Article Synopsis
- Deep learning (DL) algorithms effectively detect vascular lesions in small bowel capsule endoscopy (CE), improving diagnostic performance and reducing reading time.
- A machine learning (ML) classifier was used to enhance the DL algorithm's performance by categorizing vascular abnormalities and selecting the most relevant images for reporting.
- The random forest (RF) method achieved high specificity (91.1%) and accuracy (84.2%) in distinguishing significant lesions while dramatically reducing the number of images reported, demonstrating potential for automated CE reporting without sacrificing diagnostic quality.