Grayscale Example

OCT Image Classifier

© 2025 mRetina AI. All rights reserved. Created and designed by Ameen Marashi, M.D.

×

Instructions

System Requirements:
- Stable & Fast internet connection
- An OCT image file in a supported format (e.g., JPEG, PNG, GIF, BMP)
- PDF, DOCx and other document file types are not supported
Preparing Your OCT Image:
- Ensure the OCT image is clear and well-centered.
- Avoid images with excessive noise or artifacts.
- The image should be a grayscale ONLY like the example below.
Grayscale Example
- Avoid these types of images (shown below): Macular Map, Pseudo-color, Inverted grayscale, and captured from computer screen

Grayscale Example
Grayscale Example
Grayscale Example
Grayscale Example
Grayscale Example
Grayscale Example
Grayscale Example
Using the Classifier:
- Select the correct clinical entity.

Uploading an Image:
- Click on the designated area labeled "Drag and drop an OCT image file here, or click to select file."
- Either drag your OCT image file into the area or click to open a file selector and choose the image from your device.
- Once uploaded, the image will be displayed on the screen.

Predicting the Condition:
- Click the "Predict" button located below the displayed image; sometimes you need to re-upload the image if it doesn't appear.
- A loading overlay with a spinner will appear, indicating that the model is processing the image.
- Once the prediction is complete, the overlay will disappear, and the predicted condition will be displayed.

The image classifier is capable of predicting only specific pathologies based on the OCT images uploaded:
- The image classifier will not extract features from OCT images such as subretinal or intraretinal fluids. - When diabetic retinopathy and macular edema are selected, the OCT image classifier can predict whether the uploaded image is Central DME, Non-central CSME, DME with Vitreomacular abnormalities, or Tractional Retinal Detachment.
- If Retinal Vascular Occlusive diseases are selected, the OCT image classifier can predict whether the uploaded image is a Long-Standing Retinal Vascular Occlusion, Macular Edema Related to Retinal Vein Occlusion, or Retinal Artery Occlusion.
- When AMD is selected for the OCT image classifier, it can predict whether the uploaded image is Non-neovascular AMD, Non-neovascular AMD with Geographic atrophy, or Active Neovascular AMD.
- If Pachychoroid is selected, the OCT image classifier can predict whether the uploaded image is CSCR, CSCR complicated with MNV, or Polypoidal Choroidal Vasculopathy.
- If Vitreo Macular Interface Abnormalities are selected, the OCT image classifier can predict whether the uploaded image is Lamellar Macular Hole, Vitreomacular traction, Full-thickness Macular Hole, or Epiretinal Membrane.
- When Hereditary Retinal Diseases are selected, the OCT image classifier can predict whether the uploaded image is Retinitis Pigmentosa, Retinitis Pigmentosa with Macular Edema, X-linked Juvenile Retinoschisis, or Vitelliform Lesion.

Viewing Recommendations:
- If the predicted condition is not "Normal" a "Suggest Recommendations" button will appear.
- Click the "Suggest Recommendations" button to view treatment recommendations related to the predicted condition.
- Read the displayed recommendations carefully, but be aware that incorrect predictions will result in wrong recommendations.
- If desired, click the "Build your case with AI" button to use an AI case generator tool for further analysis.

Troubleshooting and Tips:
- No Prediction or Error: Ensure the image file is in a supported format, typed, and properly uploaded. Check your internet connection and try again.
- Inaccurate Prediction: The classifier has a 70% accuracy rate with correct images. For example, if the patient presented with cystic lesions and no increased macular thickness, then AI may falsely produce a false prediction. Ensure the image quality is optimal. Consider consulting a specialist for a professional diagnosis and recommendations." Ensure the image quality is optimal. Consider consulting a specialist for a professional diagnosis and recommendations.
- Button Not Appearing: If the "Suggest Recommendations" button does not appear, the condition might be predicted as "Normal," or there might be an issue with the prediction process. Try re-uploading the image or refreshing the page.

Disclaimer:
- This tool is intended for preliminary screening and educational purposes only.
- It should not be used as a substitute for professional medical diagnosis or treatment.
- Always consult with a qualified healthcare provider for any medical concerns or conditions.