AI Label

AI Label refers to the process of quickly segmenting and labeling image samples based on Machine Learning model.

Two Model Algorithms, PaddleSeg and SAM, are provided:

PaddleSeg: It is mainly used for image segmentation, and it contains multiple mainstream segmentation models, so as to complete image segmentation more efficiently.
SAM: Segment Anything Model, literally translated as Segmenting Everything, can segment All Objects in the picture. However, it requires higher hardware, and the video memory of the machine must be greater than 8G.

Click AI Label to pop up the AI Label dialog box;
Set the segmentation threshold. The value range is 0, 1. The default value is 0.5.
Select Model: provides two Model Algorithms, PaddleSeg and SegmentAnything. The accuracy of SegmentAnything is higher than that of PaddleSeg.
Click the label button in the AI Label dialog box to click the sample in the Sample Picture Window with the left and right mouse buttons.

Note:

clicking the label button, select the sample with the left mouse button, and end the selection of the sample with the right mouse button. At this time, the Edit Sample Tags dialog box will pop up to select the label type. After that, repeat the sample selection process and the Edit Sample Tags dialog box will no longer pop up.
A sample can be selected multiple times with the left mouse button. If the Select Range exceeds the sample range, press and hold the Ctrl key and click the left mouse button to delete the out-of-range part.