prabindh / euclid

Euclid object labeller for frictionless object detection training purposes in Machine learning frameworks (KITTI, YOLO)
45 stars 20 forks source link

Image dimensions error ... #5

Closed MyVanitar closed 7 years ago

MyVanitar commented 7 years ago

Hello,

I see this error with some images, why you put this term on input images?

2017-02-07_23-47-35

prabindh commented 7 years ago

The training phase usually involves a resize step. But it is better to resize to the typical dimensions of the framework that is being used. Typical sizes are 256x256, etc. Anything larger than a typical size is flagged with an error message as you noted above. To avoid, please resize to a smaller size.

MyVanitar commented 7 years ago

Alright, could you add a bonus tool inside your code to do this resizing (by keeping the aspect ration of course)?

besides I have these suggestions:

ip1981 commented 7 years ago

The tool exists http://www.imagemagick.org/Usage/resize/

prabindh commented 7 years ago

If we resize inside the tool, what if the user needs to control if (a) Replace the original file or (b) Create a copy and resize a new file ?

MyVanitar commented 7 years ago

That's correct.