Fetal Brain MR Segmentation

This model identifies and segments the fetal brain, if present, in an MR image. The input should be a single slice of an MR image (which is triplicated to mimic the three slices in the input training data). The output is a binary mask, where "white" is used to represent the fetal brain.

Dataset Description: This model was trained on MR images collected during the iFind (intellient Fetal Imaging and Diagnosis) project. The data was augmented using random flips in each direction and gaussian noise was added to the images as well.

Input
Output

View API [+][-]

POST /api/predict/72

Request
{
  data : string - A base64 string representation of the image to upload as input.
}

Response
{
  data : string - A base64 string representation of the image returned as output.
}