Enhancing transformer-based architectures with geometric deep learning for colonoscopic polyp size classification using transfer learning.

Accurate estimation of polyp size during colonoscopy is critical for risk assessment and surveillance planning in colorectal cancer prevention. However, current methods often rely on subjective visual judgment, leading to inconsistencies and potentia…

Preview only. Read the full abstract at the source