Configure Filtering Rules

You are currently viewing the documentation for the latest version (3.0.0). To access a different version, click the "Switch version" button located in the upper-right corner of the page.

■ If you are not sure which version of the product you are currently using, please feel free to contact Mech-Mind Technical Support.

In the Validation tab, click Filtering Rule Settings to configure filtering rules for the model. The Instance Segmentation module supports configuring Morphological transformation and Confidence Threshold as needed.

After the configuration takes effect, validation results change accordingly, and the related settings also take effect in the exported model.

Morphological Transformation

When morphological transformation is enabled, morphological processing can be applied to the mask to improve the boundary quality and connectivity of the target region.

Parameter	Description
Morphological transformation	Once enabled, morphological processing is applied to the mask.
Morphological transformation type	Four morphological operations are supported: Erosion, Dilation, Opening, and Closing. For morphological processing, the default kernel size is 3 px. Erosion: Reduces the mask area. Suitable for cases where the mask is too large or contains noise. Dilation: Increases the mask area. Suitable for cases where the mask is smaller than the actual area, to avoid missing point cloud data. Opening: Performs erosion first, then dilation on the mask. Removes small noise regions while maintaining the overall size and shape of the target. Closing: Performs dilation first, then erosion on the mask. Fills small holes while maintaining the overall size and shape of the target.

Parameter

Description

Morphological transformation

Once enabled, morphological processing is applied to the mask.

Morphological transformation type

Four morphological operations are supported: Erosion, Dilation, Opening, and Closing. For morphological processing, the default kernel size is 3 px.

Erosion: Reduces the mask area. Suitable for cases where the mask is too large or contains noise.
Dilation: Increases the mask area. Suitable for cases where the mask is smaller than the actual area, to avoid missing point cloud data.
Opening: Performs erosion first, then dilation on the mask. Removes small noise regions while maintaining the overall size and shape of the target.
Closing: Performs dilation first, then erosion on the mask. Fills small holes while maintaining the overall size and shape of the target.

Confidence Threshold

Confidence refers to the accuracy of labeling results. A higher confidence indicates more accurate labeling results. The inference result is retained only when its confidence exceeds this threshold.

You can set the confidence threshold by this step:

Set the threshold according to the actual requirement (default value: 0.4).

Setting the confidence too high may result in a lower number of displayed labels.

After completing the configuration, click OK to make the configuration take effect.

Is this page helpful?

Leave a feedback

Configure Filtering Rules

Morphological Transformation

Confidence Threshold

Is this page helpful?

Thanks for your support!

You can give a feedback in any of the following ways:

We Value Your Privacy