Your SICK ID gives you access to our extensive range of services. This includes direct online orders, price and availability checks, and access to our digital services.
In the Domains page, select the domain you created and go into the Object Finding tab.
1. Drag and drop or select your recordings folder to be uploaded to dStudio for annotation and training. You can select the entire folder, and only the 2D images will be filtered and uploaded.
2. Then create a Class for annotation or labeling.
3. And click on Annotate.
Annotate/Label your images
Annotate the data using the 'Box' tool we created, by drawing polygons along the box/object edges or by using the 'wand' tool for Assisted Object annotation. Annotate all possible boxes/objects in an images and press save. Repeat until all images are annotated.
Next step is to create a Snapshot and Project
Go back to your domain on dStudio and create a Snapshot.
After creating a Snapshot create a Project for that Snapshot.
Select if you would like to train a PALLOC or PLB network.
For PALLOC, you can train Object detection/Restricted rotation networks and Instance Segmentation/Free rotation networks for both Boxes and Sacks.
And for PLB you can train Instance segmentation networks only. The balanced networks give us better performance, and the other network give us faster segmentation results.
Select the desired network and press Next.
Give it a name and press Create Project and the network training will begin. You can keep track of the training on the Project tab on dStudio.
Once trained you can Download the full network from the Project under the Deploy tab. Download the network and load it onto PALLOC or PLB.
How to deploy the neural network once it has been trained.
To deploy a network on PALLOC, open PLxManager and select your device. Drag and drop the neural network file anywhere across the PLxManager window or right click on your device and select 'Install File on device' and select the network file.
And for PLB, in the algorithm dropdown select, dLocate Smartpick or dLocate Box. Then load the network in the Neural network path selection and it should be ready for testing!
