Data Augmentation Framework for Improving Image Recognition Using cycleGAN

Sangmin Kim, Selome Tesfaye Deribe, Gyurin Byun, Kyeongjin Joo, Jeongwon Pyo, Taeyong Kuc, Hyunseung Choo

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper proposes a novel framework that significantly enhances the performance of semantic segmentation models in recognizing specific objects. Leveraging the capabilities of Generative Adversarial Networks (GANs), particularly Cycle-GAN, this framework focuses on augmenting high-quality data to improve object recognition in autonomous driving and other applications. The study utilizes a dataset of 5,005 road images, enriched with polygon labels for precise object recognition. Key advancements in this research include the implementation of feature matching and fact forcing techniques to stabilize and integrate GAN performance, thereby overcoming common challenges like mode collapse, slow training, and overfitting. In the performance-enhanced GAN model, we improved the Discriminator Loss from the original 1.0517 to 0.0001, achieving convergence to zero 66.67% faster.

Original languageEnglish
Title of host publicationProceedings of the 2024 18th International Conference on Ubiquitous Information Management and Communication, IMCOM 2024
EditorsSukhan Lee, Hyunseung Choo, Roslan Ismail
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350331011
DOIs
StatePublished - 2024
Event18th International Conference on Ubiquitous Information Management and Communication, IMCOM 2024 - Kuala Lumpur, Malaysia
Duration: 3 Jan 20245 Jan 2024

Publication series

NameProceedings of the 2024 18th International Conference on Ubiquitous Information Management and Communication, IMCOM 2024

Conference

Conference18th International Conference on Ubiquitous Information Management and Communication, IMCOM 2024
Country/TerritoryMalaysia
CityKuala Lumpur
Period3/01/245/01/24

Keywords

  • Computer Vision
  • Generative Adversarial Networks
  • Optimization modelling
  • Semantic Segmentation

Fingerprint

Dive into the research topics of 'Data Augmentation Framework for Improving Image Recognition Using cycleGAN'. Together they form a unique fingerprint.

Cite this