# Latent Filter Scaling for Multimodal Unsupervised Image-to-Image Translation

arXiv: https://arxiv.org/abs/1812.09877 · HTML: https://mlsanity.com/paper/1812.09877/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 24 Dec 2018.

## Built on by (most cited first, 10 of 10)

- [Image-to-Image Translation: Methods and Applications](https://mlsanity.com/paper/2101.08629/) (2021)
- [ModDrop++: A Dynamic Filter Network with Intra-subject Co-training for Multiple Sclerosis Lesion Segmentation with Missing Modalities](https://mlsanity.com/paper/2203.04959/) (2022)
- [MicroAST: Towards Super-Fast Ultra-Resolution Arbitrary Style Transfer](https://mlsanity.com/paper/2211.15313/) (2022)
- [A Unified Hyper-GAN Model for Unpaired Multi-contrast MR Image Translation](https://mlsanity.com/paper/2107.11945/) (2021)
- [Adversarial Continual Learning for Multi-Domain Hippocampal Segmentation](https://mlsanity.com/paper/2107.08751/) (2021)
- [Controllable cardiac synthesis via disentangled anatomy arithmetic](https://mlsanity.com/paper/2107.01748/) (2021)
- [Multi-scale Super-resolution Magnetic Resonance Spectroscopic Imaging with Adjustable Sharpness](https://mlsanity.com/paper/2206.08984/) (2022)
- [Heredity-aware Child Face Image Generation with Latent Space Disentanglement](https://mlsanity.com/paper/2108.11080/) (2021)
- [SuperStyleNet: Deep Image Synthesis with Superpixel Based Style Encoder](https://mlsanity.com/paper/2112.09367/) (2021)
- [Referenceless User Controllable Semantic Image Synthesis](https://mlsanity.com/paper/2306.10646/) (2023)

## Abstract (arXiv)

In multimodal unsupervised image-to-image translation tasks, the goal is to translate an image from the source domain to many images in the target domain. We present a simple method that produces higher quality images than current state-of-the-art while maintaining the same amount of multimodal diversity. Previous methods follow the unconditional approach of trying to map the latent code directly to a full-size image. This leads to complicated network architectures with several introduced hyperparameters to tune. By treating the latent code as a modifier of the convolutional filters, we produce multimodal output while maintaining the traditional Generative Adversarial Network (GAN) loss and without additional hyperparameters. The only tuning required by our method controls the tradeoff between variability and quality of generated images. Furthermore, we achieve disentanglement between source domain content and target domain style for free as a by-product of our formulation. We perform qualitative and quantitative experiments showing the advantages of our method compared with the state-of-the art on multiple benchmark image-to-image translation datasets.