Darknet: number of images produced by data augmentation

Created on 11 Jul 2019  路  3Comments  路  Source: AlexeyAB/darknet

Hi @AlexeyAB
I wanted to know how many images are produced by auto-data augmentation in yolo-cfg

picture jasleen137

Most helpful comment

I believe that every image the network every sees is essentially unique, generated with randomised parameters (within a fairly well defined param space)

picture LukeAI on 11 Jul 2019
👍2

All 3 comments

I believe that every image the network every sees is essentially unique, generated with randomised parameters (within a fairly well defined param space)

LukeAI picture LukeAI on 11 Jul 2019
👍2

@LukeAI Yes.

AlexeyAB picture AlexeyAB on 11 Jul 2019

@AlexeyAB could you please provide information what types of augmentation are applied during the training procedure? For instance, in xview-yolov3 they apply the following tasks

Augmentation | Description
--- | ---
Translation | +/- 1% (vertical and horizontal)
Rotation | +/- 20 degrees
Shear | +/- 3 degrees (vertical and horizontal)
Scale | +/- 30%
Reflection | 50% probability (vertical and horizontal)
HSV Saturation | +/- 50%
HSV Intensity | +/- 50%

I would like understand and be 100% sure what is done during this procedure.

BERENZ picture BERENZ on 28 Aug 2019
👍1
Was this page helpful?
0 / 5 - 0 ratings

Related issues

shootingliu picture shootingliu  路  3Comments
kebundsc picture kebundsc  路  3Comments
off99555 picture off99555  路  3Comments
zihaozhang9 picture zihaozhang9  路  3Comments
Mididou picture Mididou  路  3Comments