GANiry: Bald-to-Hairy Translation Using CycleGAN
Fidan Samet, Oguz Bakir
TL;DR
This paper introduces GANiry, a CycleGAN-based model that translates bald men to hairy men with realistic, stylized, and colored hair, using perceptual loss and conditional constraints for enhanced results.
Contribution
The paper extends CycleGAN with perceptual loss and conditional constraints to improve hair translation quality and style diversity in bald-to-hairy men images.
Findings
Perceptual loss improves realism of generated hair.
Conditional constraints enable diverse hair styles and colors.
Qualitative results demonstrate effective hair translation.
Abstract
This work presents our computer vision course project called bald men-to-hairy men translation using CycleGAN. On top of CycleGAN architecture, we utilize perceptual loss in order to achieve more realistic results. We also integrate conditional constrains to obtain different stylized and colored hairs on bald men. We conducted extensive experiments and present qualitative results in this paper. Our code and models are available at https://github.com/fidansamet/GANiry.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsGenerative Adversarial Networks and Image Synthesis · Advanced Vision and Imaging · Computer Graphics and Visualization Techniques
Methods*Communicated@Fast*How Do I Communicate to Expedia? · Batch Normalization · Residual Connection · Instance Normalization · Residual Block · Sigmoid Activation · Convolution · Tanh Activation · Cycle Consistency Loss · HuMan(Expedia)||How do I get a human at Expedia?