Multi-target in Albumentations
Many images, many masks, bounding boxes, and key points. How to transform them in sync?
data-augmentation image-augmentation computer-vision albumentations multi-target article library

Albumentations work the best with the standard tasks of classification, segmentation, object, and keypoint detection. But there are situations when your samples consist of a set of different objects.

  • Multi-target functionality specifically designed for this situation.
  • Possible use cases.
  • Siamese networks
  • Sequential frames in the video.
  • Image2image.
  • Multilabel segmentation.
  • Instance segmentation.
  • Panoptic segmentation.

Don't forget to tag @ternaus in your comment, otherwise they may not be notified.

Authors community post
Computer Vision for Autonomous Vehicles at Lyft. PhD in Physics from UC Davis, Kaggle GrandMaster
Share this project
Similar projects
Image Augmentations for GAN Training
We systematically study the effectiveness of various existing augmentation techniques for GAN training in a variety of settings.
Modern Data Augmentation Techniques for Computer Vision
A bunch of modern data augmentation techniques for computer vision covering cutout, mixup, cutmix and augmix.
Cut, Paste and Learn: Surprisingly Easy Synthesis for Detection
Generate synthetic scenes and bounding box annotations for object detection.
Classify photos in 600 classes using nine million Open Images
If you’re looking build an image classifier but need training data, look no further than Google Open Images.