Asked
— Edited
Is it possible to change the size of the capture red box (80x80)? If i change the size from 320 to 640 camera resolution, the box is so small I have to be across the room to train. Am I remembering incorrectly, wasn't the box at one point sizable, a stretchy box?
No - the algorithm supports the size provided. It's recommended to follow the tutorial via The Robot Program to understand how to train an object. For example, it's not necessary to train the entire object - instead train specific characteristic details of the object. Training the object will include the background - and the background is not part of the object. It should not be assumed that the software knows the difference between the background of a room AND the object.
Thanks DJ,
What seems to work best is to train on the lowest camera setting and then change the settings to 640 (for a clearer image), still tracks just fine.
Is it possible that the "blue" tracking box be as bold (thickness of line) on the 640 or higher settings, as it is on the 160 resolution settings. I will be building out a display that will show on a large monitor. The bolder the better.
I remember you chatting with someone else on a thread about this so i will add to my results here. I'm having the same issue with the training window control not showing live on the red portion of my screen, it freezes. It is just the object training window. No big deal for me just adding my data.
Windows 10 Nvidia GTX 1080 TI 12 gigs v ram
EDIT: Camera control window when dragged around does not do it, just training window.