Labeling

another big advantages of trescope is labeling, for now, we support image classification labeling, point classification labeling, mesh classification labeling, we will support pointcloud semantic labeling, stay tuned.

classification labeling

trescope provides efficient interfaces for labeling whenever need, for example, one need to label the leg attribute in image, total attributes are: circle leg, square leg, another leg attribute is size, total attributes are: xs, s, m, l, xl . trescope is easy to accompany this task, use the following command to plot them

Trescope().selectOutput(sample_index * 2).plotImage(full_image_path).withConfig(ImageConfig())
 (Trescope().selectOutput(sample_index * 2 + 1).asInput()
     .addControl(Label().id('file').label(model_image_path).value(full_image_path))
     .addControl(EnumControl().id('0').label('0').enumeration(0, 1).defaultValue(0))
     .addControl(EnumControl().id('shape').label('Circle or Square').enumeration('circle', 'square').style('vertical'))
     .addControl(EnumControl().id('thickness').label('Size').enumeration('xs', 's', 'm', 'l', 'xl').defaultValue('m'))
     .addControl(TextControl().id('otherInfo').label('OtherInfo').defaultValue('something'))
     )

firstly, plot image or pointcloud or mesh in selected grid, secondly, using Label() to name the image path or something related, the key part is EnumControl() , id() specifies the name, label() specifies the meaning shown in grid for labeling, EnumControl() specifies all the attributes, btw, this is not required, one can set the default value using defaultValue(‘something’) .

pose labeling

we will support object pose labeling, stay tuned.