Process Neural Network (Inferencing)
At runtime, each sample from the runtime image is processed individually, using the tool's trained network, and an individual network response is obtained for each sample. The response from the network is expressed as a probability map, where each pixel in the sampled region of the input image is assigned a probability. The meaning of the probability map depends on which tool is being used. For the Green Classify tool, the response is the classification and the probability that the classification is correct.
These whole-image probability maps are assembled from the interpolated network responses to individual samples. The final results returned to the user (feature poses and identities, and defect regions) are based on a results formation process that the user controls by specifying thresholds for class probability.
In the above context, it is important to understand that the probability returned by a tool may not reflect well with our notion of how likely a certain judgment is. This is mostly due to the fact that the tools have a "limited view of the world" in that they do not return probabilities with respect to our wide and rich visual experience, but rather with respect to their very limited visual world of a couple of classes. Thus, for instance, in the case of a Green Classify tool, it is very well possible that the tool will assign a very high probability to a new and therefore unknown class even though these classes look very dissimilar to us. This is simply because to the tool that class looks many times more similar to that known class than any other.
The processing (inferencing) of a trained tool is automatically executed when the training is finished without issues. If you want to manually re-process the trained tool, click 
button.
Configure Processing Parameters
The Processing parameters control the way that images are processed by the tool. This is often called ‘inference’ in deep learning. Processing with the same models will always give you the same results. Changing these parameters does not require the tool to be retrained; the effect can be seen right away by reprocessing the database. To re-process the tool, click button. There are two processing parameters in High Detail mode.
| Parameters | Description |
|---|---|
| Threshold |
The threshold parameter determines the minimum score of a view to be classified.
Note: The default threshold for High Detail mode is 0% because it calculates the score for each tag in a different way from Focused mode.
It is not usually recommended to adjust threshold value in High Detail mode. |
| Heat map |
The “Heat map” in processing parameter is an option that allows user to check which area have been used as a clue to High Detail mode. You can use it as a debugging tool after generating the model with given data and ground truth tags(labels). If the Heat map sees different area with the regions that are critical to classification, it might be better to train again to generate new model. Heat map is not an accurate tool. The location of the red area could be different with what human think, so please refer it only for checking the tendency. Note: The size of workspace could be bigger if you use "heatmap" option in High Detail mode.
This helps to find the clue inside the image, but it generated heat maps that consumes lots of storage
.
To make workspace smaller, process without heatmap again and save the workspace. Green Classify High Detail projects heat maps directly onto the image viewer when you enable both the Heat Map checkbox in Processing Parameter and the one in Overlay Checkbox. |