More Augmentations

[jaideep11061982]

❔Question

How i can use more and more augmentations related to Image Enhancements.

Additional context

[github-actions[bot]]

👋 Hello @jaideep11061982, thank you for your interest in 🚀 YOLOv5! Please visit our ⭐️ Tutorials to get started, where you can find quickstart guides for simple tasks like Custom Data Training all the way to advanced concepts like Hyperparameter Evolution.

If this is a 🐛 Bug Report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you.

If this is a custom training ❓ Question, please provide as much information as possible, including dataset images, training logs, screenshots, and a public link to online W&B logging if available.

For business inquiries or professional support requests please visit https://www.ultralytics.com or email Glenn Jocher at glenn.jocher@ultralytics.com.

Requirements

Python 3.8 or later with all requirements.txt dependencies installed, including torch>=1.7. To install run:

$ pip install -r requirements.txt

Environments

YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

• Google Colab and Kaggle notebooks with free GPU:
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Amazon Deep Learning AMI. See AWS Quickstart Guide
• Docker Image. See Docker Quickstart Guide

Status

If this badge is green, all YOLOv5 GitHub Actions Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training (train.py), testing (test.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu every 24 hours and on every commit.

[glenn-jocher]

@jaideep11061982 what augmentations are you interested in?

[tetelevm]

@glenn-jocher I am also interested in this question. As I understand it, YOLO slices images for training, but does it also know how to rotate images by (90, 180, 270) degrees? (And is it even worth doing that?)

[glenn-jocher]

@tetelevm hyperparameters for augmentation including rotation are defined in the hyp file: https://github.com/ultralytics/yolov5/blob/fbf41e09134b113f8e79ae01b4eee40d00797b2d/data/hyp.scratch.yaml#L22-L33

[decent-engineer-decent-datascientist]

@glenn-jocher have you checked this out: https://github.com/facebookresearch/AugLy

I'd never need this for my use cases, but you can do stuff like add ig filters, emojis, simulate screenshots, etc. A lot of modern digital-life augmentations that might not win you any AP on COCO, but would be useful for real-world applications.

Here's a list of augmentations: https://github.com/facebookresearch/AugLy/blob/main/augly/image/__init__.py

[glenn-jocher]

@decent-engineer-decent-datascientist yes I've seen the facebook augmentation package, which seems to be the flavor of the week, but I would recommend Albumentations: https://github.com/albumentations-team/albumentations

You could integrate additional Albumentations augmentations in the YOLOv5 dataloader here: https://github.com/ultralytics/yolov5/blob/e8810a53e83ddb5dd6bf8e871c2ede701007047c/utils/datasets.py#L536-L552