glenn-jocher
commented
2 years ago @jaiswati 👋 Hello! Thanks for asking about image augmentation. YOLOv5 🚀 applies online imagespace and colorspace augmentations in the trainloader (but not the testloader) to present a new and unique augmented Mosaic (original image + 3 random images) each time an image is loaded for training. Images are never presented twice in the same way.
The hyperparameters used to define these augmentations are in your hyperparameter file (default data/hyp.scratch.yaml) defined when training:
python train.py --hyp hyp.scratch.yamlYou can view the effect of your augmentation policy in your train_batch*.jpg images once training starts. These images will be in your train logging directory, typically yolov5/runs/train/exp:
train_batch0.jpg shows train batch 0 mosaics and labels:
Good luck and let us know if you have any other questions!
github-actions[bot]
Hi @glenn-jocher
Thank you for the responses !!
is that advisable to modify the COCO format to update the coordinates of the object with a python script, as the original training image is rotated with OpenCV python script, so the corresponding position of the bb-box and mask has to be rotated? or is there any alternative to this like rotation augmentation etc. that automatically takes care of rotating bb-box/masks etc? Actually, In some of the training images, on circular ring numbers and characters are printed, we found that if numbers/characters are rotated(tilted), then model detection accuracy is less, vs when numbers/characters are not rotated (tilted) then model detection accuracy is more, so we are thinking to change the training images and modifying corresponding box/mask co-ordinates in COCO format. Best Regards, @jaiswati