Visual Turing Test
Regular OCT vs AI-Enhanced OCT Images
We have developed several deep learning models to enhance low-quality vitreous OCT images. Your clinical expertise is invaluable in helping us determine which models produce the most clinically relevant results.
In this test, you will be shown 10 low-quality vitreous OCT images. For each image, you will see 6 enhanced versions created by different AI models. Your task is to rank these enhanced images from best to worst based on your clinical judgment.
The test will take approximately 10-15 minutes to complete. Your responses will be anonymized and used solely for research purposes.
On the left, the true low-quality OCT image (ART10) acquired with the Heidelberg OCT Spectralis.
On the right, one of the enhanced versions obtained from one of our AI models.
Important Notice:
Please take this test using a desktop/laptop computer with high brightness. Mobile or tablet screens may not properly display the OCT images, which could result in inaccurate evaluations.