Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
[CVPR22] Official Implementation of DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentation
| Date | Stars |
|---|---|
| 2026-07-24 | 571 |
| 2026-07-25 | 571 |
| 2026-07-28 | 571 |
| 2026-07-30 | 571 |
| 2026-08-07 | 572 |
| 2026-08-15 | 573 |
| 2026-08-26 | 574 |
| 2026-09-20 | 574 |
Today
— stars today
This week
— stars this week
This month
+1 stars this month
Momentum
0.0
growth rate 0.00%/day
## DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentation
**by [Lukas Hoyer](https://lhoyer.github.io/), [Dengxin Dai](https://vas.mpi-inf.mpg.de/dengxin/), and [Luc Van Gool](https://scholar.google.de/citations?user=TwMib_QAAAAJ&hl=en)**
**[[CVPR22 Paper]](https://arxiv.org/pdf/2111.14887.pdf)**
**[[Extension Paper]](https://arxiv.org/pdf/2304.13615.pdf)**
:bell: **News:**
* [2024-07-03] We are happy to announce that our work [SemiVL](https://github.com/google-research/semivl) on semi-supervised semantic segmentation with vision-language guidance was accepted at **ECCV24**.
* [2024-07-03] We are happy to announce that our follow-up work [DGInStyle](https://dginstyle.github.io/) on image diffusion for domain-generalizable semantic segmentation was accepted at **ECCV24**.
* [2023-09-26] We are happy to announce that our [Extension Paper](https://arxiv.org/pdf/2304.13615.pdf) on domain generalization and clear-to-adverse-weather UDA was accapted at **PAMI**.
* [2023-08-25] We are happy to announce that our follow-up work [EDAPS](https://github.com/susaha/edaps) on panoptic segmentation UDA was accepted at **ICCV23**.
* [2023-04-23] We further extend DAFormer to domain generalization and clear-to-adverse-weather UDA in the [Extension Paper](https://arxiv.org/pdf/2304.13615.pdf).
* [2023-02-28] We are happy to announce that our follow-up work [MIC](https://github.com/lhoyer/MIC) on context-enhanced UDA was accepted at **CVPR23**.
* [2022-07-06] We are happy to announce that our follow-up work [HRDA](https://github.com/lhoyer/HRDA) on high-resolution UDA was accepted at **ECCV22**.
* [2022-03-09] We are happy to announce that DAFormer was accepted at **CVPR22**.
## Overview
As acquiring pixel-wise annotations of real-world images for semantic
segmentation is a costly process, a model can instead be trained with more
accessible synthetic data and adapted to real images without requiring their
annotations. This process is studied in **Unsupervised Domain Adaptation (UDA)**.
Even though a large number of methods propose new UDA strategies, they
are mostly based on outdated network architectures. In this work, we
particularly study the influence of the network architecture on UDA performance
and propose **DAFormer**, a network architecture tailored for UDA. It consists of a
Transformer encoder and a multi-level context-aware feature fusion decoder.
DAFormer is enabled by three simple but crucial training strategies to stabilize the
training and to avoid overfitting the source domain: While the
**Rare Class Sampling** on the source domain improves the quality of pseudo-labels
by mitigating the confirmation bias of self-training towards common classes,
the **Thing-Class ImageNet Feature Distance** and a **Learning Rate Warmup** promote
feature transfer from ImageNet pretraining.
DAFormer significantly improves
the state-of-the-art performance **by 10.8 mIoU for GTA→Cityscapes**
and **by 5.4 mIoU for Synthia→Cityscapes** and enables learning even
difficult classes such as train, bus, and truck well.
The strengths of DAFormer, compared to the previous state-of-the-art UDA method
ProDA, can also be observed in qualitative examples from the Cityscapes
validation set.
DAFormer can be further **extended to domain generalization** lifting the requirement
of access to target images. Also in domain generalization,
DAFormer significantly improves the state-of-the-art performance by **+6.5 mIoU**.
For more information on DAFormer, please check our
[[CVPR Paper]](https://arxiv.org/pdf/2111.14887.pdf) and the [[Extension Paper]](https://arxiv.org/pdf/2304.13615.pdf).
If you find this project useful in your research, please consider citing:
```
@InProceedings{hoyer2022daformer,
title={{DAFormer}: Improving Network Architectures and Training Strategies for Domain-Excerpt of 18,009 characters
Read on GitHub
21
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:4b83033da4870417, topic:transformer