Hello,
I’m encountering an issue with one of my segmentation models (dental_project-xcawb/1) when using the Roboflow Inference API in my Streamlit application. The model is designed to segment Teeth anomalies.
Project Type: Segmentation, Classification (Disease Identification)
Operating System & Browser:MacOs/ Google Chrome
Setup:
API Endpoint:I am using the serverless inference endpoint: https://serverless.roboflow.com.
Inference SDK:I am using the inference_sdk in Python.
Authentication: I am using my correct API key.
Classification Model:I am also using a classification model (sinistroodonto/1) which correctly identifies the images as containing a mouth ulcer with high confidence (e.g., 87.7%).
Code Snippet:
Observed Behavior:
The classification API call correctly identifies the “Mouth Ulcer”.
The segmentation API call completes without errors.
The Raw Segmentation Result consistently shows 'predictions': [].
This happens across multiple different images of mouth ulcers that work correctly in the Roboflow web interface tester for the same model.
My Question:
Has anyone else experienced a similar issue where a segmentation model works in the Roboflow web interface but returns empty predictions via the API? Are there any specific considerations or parameters I might be missing when making the API call for segmentation? Any insights or suggestions on how to resolve this would be greatly appreciated.
Thank you for your time and help!