-
Hey!
The new feature that allows using the same augmentations for the images and their corresponding ground-truth is great, I am trying to use it with the pytorch transform.
The problem is that the …
-
I am working on a image segmentation problem where I would like to augment images with a condition, that all color augmentation methods does not affect ground truth data
Is there a way where I can …
-
Hi all, this might be a basic question, but in this [line](https://github.com/gidariss/FeatureLearningRotNet/blob/master/dataloader.py#L106), and other lines in the dataloader, the `RandomHorizontalFl…
-
### Search before asking
- [x] I have searched the Ultralytics YOLO [issues](https://github.com/ultralytics/ultralytics/issues) and [discussions](https://github.com/ultralytics/ultralytics/discussion…
-
**User Feedback:**
Keep in mind, some of these augmentation tasks will need to be applied to the ground truth image labels as well (e.g. masks). Mainly the ones that change the scale, occlude, or ori…
-
In the data augmentation part of the paper,
"The size of each sampled patch is [0.1, 1] of the original image size, and the aspect ratio
is between 1/2and 2. We keep the overlapped part of the grou…
-
Hi Hoel,
Thanks very much for posting your work on the boundary loss! I really appreciate what you've done. However I have a question with regards to data augmentation. I am currently working on…
-
I couldn't locate the data augmentation preprocessing mentioned in the paper's algorithm, specifically in Algorithm 2, which is designed to take ground truth samples as input and generate augmented sa…
-
Hi @dbolya ,
I just used basic configuration to train on my own dataset (which is in coco format), and I only got around 20 MAP in both box and mask perspectives (30+ epochs). Also, the speed test …
-
Hello,
Actually working on a multimodal pixel wise segmentation, my dataset is composed of multi inputs with only one ground truth (times N_images).
From that situations, the ideal case would be t…