ANSTS supports a built-in test interface to allow users to test the model inference before final deployment.
To start testing the model, first you need to choose the model type (GPU/CPU)
Then, you can select the image source for testing:
- Current project images: some of the images from the training dataset will be displayed
- Other image sources: you can browse to other image folders locally
- Webcam: Use a webcam or integrated camera for real-time inferencing
On the right-side panel, you can:
- Control how the image is displayed
- Adjust the min score (detection threshold)
- Review the inference speed (object detection time), and detected objects in the current image
- Filter the object you want to see.