vietanhdev / anylabeling

Effortless AI-assisted data labeling with AI support from YOLO, Segment Anything (SAM+SAM2), MobileSAM!!
https://anylabeling.nrl.ai
GNU General Public License v3.0
2.27k stars 240 forks source link

Support sliders and inputting scaling factors to uniformly scale the automatically annotated polygons or rectangular bounding boxes. #91

Open aiyou9 opened 1 year ago

aiyou9 commented 1 year ago

When using a low-precision model to annotate objects with unclear edges or shadows, the automatically generated annotation boxes often fall mostly or entirely within the interior of the object's edges. Manual adjustments are then required. To minimize the manual adjustments, a feature for expanding the annotation boxes with a single click in sliders and inputting scaling factors can be useful. After expanding the annotation boxes, the refinement functionality mentioned in issue #90 can be used for further adjustments.

当使用低精度的模型,进行边缘不清晰或有阴影的目标时,自动标注产生的标注框,往往整体或大部分处于目标边缘的内侧,需要手动调整,为了减少手动调整,需要使用滑块或输入放大比例系数,一键扩大标注框,然后再使用 #90 所述功能,进行调整

The slider can toggle between different orders of magnitude for fine adjustments, with incremental steps of 1, 0.1, 0.01, and 0.001.

滑块可以切换不同数量级的步长,进行精细调整,就是在1,0.1,0.01,0.001的递增步长