Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AugmenTory: A Fast and Flexible Polygon Augmentation Library

Tanaz Ghahremani, Mohammad Hoseyni, Mohammad Javad Ahmadi, Pouria Mehrabi, Amirhossein Nikoofard

TL;DR

The paper addresses the data scarcity challenge in instance segmentation by focusing on polygon augmentations, a task poorly supported by existing libraries. It proposes AugmenTory, a keypoint-based approach that applies Albumentations transforms to polygon vertices, reconstructs polygons, and uses an IoU-based threshold to filter invalid or low-overlap augmentations. The method achieves substantial improvements in time and memory efficiency over traditional mask-based augmentation, and includes a post-processing thresholding mechanism to control object overlap during augmentation. Its design emphasizes flexibility and compatibility with common deep learning frameworks, making polygon augmentation more scalable for practical computer vision workflows.

Abstract

Data augmentation is a key technique for addressing the challenge of limited datasets, which have become a major component in the training procedures of image processing. Techniques such as geometric transformations and color space adjustments have been thoroughly tested for their ability to artificially expand training datasets and generate semi-realistic data for training purposes. Data augmentation is the most important key to addressing the challenge of limited datasets, which have become a major component of image processing training procedures. Data augmentation techniques, such as geometric transformations and color space adjustments, are thoroughly tested for their ability to artificially expand training datasets and generate semi-realistic data for training purposes. Polygons play a crucial role in instance segmentation and have seen a surge in use across advanced models, such as YOLOv8. Despite their growing popularity, the lack of specialized libraries hampers the polygon-augmentation process. This paper introduces a novel solution to this challenge, embodied in the newly developed AugmenTory library. Notably, AugmenTory offers reduced computational demands in both time and space compared to existing methods. Additionally, the library includes a postprocessing thresholding feature. The AugmenTory package is publicly available on GitHub, where interested users can access the source code: https://github.com/Smartory/AugmenTory

AugmenTory: A Fast and Flexible Polygon Augmentation Library

TL;DR

The paper addresses the data scarcity challenge in instance segmentation by focusing on polygon augmentations, a task poorly supported by existing libraries. It proposes AugmenTory, a keypoint-based approach that applies Albumentations transforms to polygon vertices, reconstructs polygons, and uses an IoU-based threshold to filter invalid or low-overlap augmentations. The method achieves substantial improvements in time and memory efficiency over traditional mask-based augmentation, and includes a post-processing thresholding mechanism to control object overlap during augmentation. Its design emphasizes flexibility and compatibility with common deep learning frameworks, making polygon augmentation more scalable for practical computer vision workflows.

Abstract

Data augmentation is a key technique for addressing the challenge of limited datasets, which have become a major component in the training procedures of image processing. Techniques such as geometric transformations and color space adjustments have been thoroughly tested for their ability to artificially expand training datasets and generate semi-realistic data for training purposes. Data augmentation is the most important key to addressing the challenge of limited datasets, which have become a major component of image processing training procedures. Data augmentation techniques, such as geometric transformations and color space adjustments, are thoroughly tested for their ability to artificially expand training datasets and generate semi-realistic data for training purposes. Polygons play a crucial role in instance segmentation and have seen a surge in use across advanced models, such as YOLOv8. Despite their growing popularity, the lack of specialized libraries hampers the polygon-augmentation process. This paper introduces a novel solution to this challenge, embodied in the newly developed AugmenTory library. Notably, AugmenTory offers reduced computational demands in both time and space compared to existing methods. Additionally, the library includes a postprocessing thresholding feature. The AugmenTory package is publicly available on GitHub, where interested users can access the source code: https://github.com/Smartory/AugmenTory
Paper Structure (4 sections, 2 figures, 1 table, 1 algorithm)

This paper contains 4 sections, 2 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Methodology
  3. Experiment
  4. Discussion and Conclusion

Figures (2)

  • Figure 1: The experiment on COCO dataset with the threshold of 0
  • Figure 2: The experiment of COCO dataset while \ref{['fig: HorseCrop0']} has 0% overlap and \ref{['fig: HorseCrop20']} has 20% overlap for post-process threshold.