The WebApp will score the degree of artifact in the 3D OCTA volumes into “Good” or “Suboptimal”. This is based on a neural network algorithm, OCTA-GAN.
Usage: Drag and drop 3mm x 3mm Zeiss OCTA .img files into the gray box. Depending on the computer you could select up to approximately 30 at a time.
Output Data
References:
- Sumpena E, Kashani AH, Jones C, “Unsupervised Artifact Detection in 3D OCTA Microvascular Imaging based on Generative Adversarial Networks”, Annual ARVO meeting in Seattle WA, Poster A0216, May 6, 2024.
- Jones C, Sumpena E, Cornelio A, Seshadri S, Beiser A, Kashani AH, “Quantitative Study on the Generalizability of 3D OCTA Artifact Detection with OCTA-GAN”, Annual ARVO meeting in Salt Lake City UT, Poster A107, May 5, 2025.