{
  "inputs": [
    {
      "image_id": "image-0",
      "image_base64": "REPLACE_WITH_BASE64_PNG_OR_JPEG"
    }
  ],
  "task": "depth_estimation",
  "parameters": {}
}