論文速覽：Multi-source Domain Adaptation for Semantic Segmentation
Multi-source Domain Adaptation for Semantic Segmentation
[NeurIPS 2019] [2020: MADAN: Multi-source Adversarial Domain Aggregation Network for Domain Adaptation] [github]
目錄
Multi-source Domain Adaptation for Semantic Segmentation
Dynamic Adversarial Image Generation
Adversarial Domain Aggregation
Feature-aligned Semantic Segmentation
Abstract
Simulation-to-real domain adaptation for semantic segmentation has been actively studied for various applications such as autonomous driving. Existing methods mainly focus on a single-source setting, which cannot easily handle a more practical scenario of multiple sources with different distributions. In this paper, we propose to investigate multi-source domain adaptation for semantic segmentation. Specifically, we design a novel framework, termed Multi-source Adversarial Domain Aggregation Network (MADAN), which can be trained in an end-to-end manner. First, we generate an adapted domain for each source with dynamic semantic consistency while aligning at the pixel-level cycle-consistently towards the target. Second, we propose sub-domain aggregation discriminator and cross-domain cycle discriminator to make different adapted domains more closely aggregated. Finally, feature-level alignment is performed between the aggregated domain and target domain while training the segmentation network. Extensive experiments from synthetic GTA and SYNTHIA to real Cityscapes and BDDS datasets demonstrate that the proposed MADAN model outperforms state-of-the-art approaches.
第一句，研究意義：在自動駕駛等多種應用中，針對語義分割的模擬-真實域適應問題進行了研究。
第二句，提出問題：現有的方法主要集中在單一源域設定上，這很難處理具有不同分佈的、多個源的、更實際的場景。
第三句，本文工作：一句話概括本文做的內容，即 提出研究語義分割的多源域適應問題。
第四-七句，具體方法：本文提出的模型，Multi-source Adversarial Domain Aggregation Network (MADAN)，包括三個部分：
First，為每個源生成一個具有動態語義一致性的自適應域，同時在畫素級迴圈上對齊目標；
Second，出了子域聚集判別器和跨域迴圈判別器，以使不同的適應域更緊密地聚集在一起；
Finally，在訓練分割網路的同時，對聚集的域和目標域進行特徵級對齊。
第八句，實驗結論：從合成的資料集 GTA 和 SYNTHIA 到真實的資料集 Cityscapes 和 BDDS 上的廣泛實驗表明，所提出的 MADAN 模型優於最先進的方法。
Problem Setup
MADAN
Overview
Figure 1: The framework of the proposed Multi-source Adversarial Domain Aggregation Network (MADAN). The colored solid arrows represent generators, while the black solid arrows indicate the segmentation network F. The dashed arrows correspond to different losses.
Dynamic Adversarial Image Generation
Adversarial Domain Aggregation
Feature-aligned Semantic Segmentation
MADAN Learning
相關文章
- 【論文速遞】PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic SegmentationCloudSegmentation
- 論文解讀（CAN）《Contrastive Adaptation Network for Unsupervised Domain Adaptation》ASTAPTAI
- 論文解讀（CDCL）《Cross-domain Contrastive Learning for Unsupervised Domain Adaptation》ROSAIASTAPT
- 論文解讀（PCL）《Probabilistic Contrastive Learning for Domain Adaptation》ASTAIAPT
- 論文閱讀筆記：Fully Convolutional Networks for Semantic Segmentation筆記Segmentation
- Heterogeneous Domain Adaptation 異質域適應論文合集AIAPT
- 論文解讀（SGDA）《Semi-supervised Domain Adaptation in Graph Transfer Learning》AIAPT
- [論文][半監督語義分割]Adversarial Learning for Semi-Supervised Semantic SegmentationSegmentation
- 深度學習論文翻譯解析（八）：Rich feature hierarchies for accurate object detection and semantic segmentation深度學習ObjectSegmentation
- [論文][半監督語義分割]Semi-Supervised Semantic Segmentation with High- and Low-level ConsistencySegmentation
- [論文速覽] Learning to Write Stylized Chinese CharactersZed
- [論文速覽] Separating Style and Content for Generalized Style TransferZed
- 遷移學習（DCCL）《Domain Confused Contrastive Learning for Unsupervised Domain Adaptation》遷移學習AIASTAPT
- 遷移學習（ADDA）《Adversarial Discriminative Domain Adaptation》遷移學習AIAPT
- 遷移學習（JDDA） 《Joint domain alignment and discriminative feature learning for unsupervised deep domain adaptation》遷移學習AIAPT
- 遷移學習（IIMT）——《Improve Unsupervised Domain Adaptation with Mixup Training》遷移學習AIAPT
- 論文閱讀-Causality Inspired Representation Learning for Domain GeneralizationAI
- [論文速覽] CalliGAN@ Style and Structure-aware Chinese Calligraphy Character GeneratorStruct
- [論文速覽] Design and Development of a Framework For Stroke-Based Handwritten Gujarati Font GenerationdevFrameworkJAR
- 遷移學習《Efficient and Robust Pseudo-Labeling for Unsupervised Domain Adaptation》遷移學習AIAPT
- 一文速覽 Dubbo 3.0
- 虛假新聞檢測（CADM）《Unsupervised Domain Adaptation for COVID-19 Information Service with Contrastive Adversarial Domain Mixup》AIAPTORMAST
- 論文速讀紀錄 - 202408
- 論文速讀記錄 - 202409
- 遷移學習（CLDA）《CLDA: Contrastive Learning for Semi-Supervised Domain Adaptation》遷移學習LDAASTAIAPT
- 論文解讀（MLDG）《Learning to Generalize: Meta-Learning for Domain Generalization》AI
- 論文閱讀：2023_Semantic Hearing: Programming Acoustic Scenes with Binaural Hearables
- 一文速覽React全棧React全棧
- 並行多工學習論文閱讀（一）：多工學習速覽並行
- 遷移學習（COAL）《Generalized Domain Adaptation with Covariate and Label Shift CO-ALignment》遷移學習ZedAIAPT
- 遷移學習（PAT）《Pairwise Adversarial Training for Unsupervised Class-imbalanced Domain Adaptation》遷移學習AIAPT
- 【機器學習】李宏毅——Domain Adaptation(領域自適應)機器學習AIAPT
- 【論文筆記】Cross-Domain WiFi Sensing with Channel State Information: A Survey筆記ROSAIWiFiORM
- 論文翻譯：2021_Acoustic Echo Cancellation with Cross-Domain LearningROSAI
- 遷移學習（TSRP）《Improving Pseudo Labels With Intra-Class Similarity for Unsupervised Domain Adaptation》遷移學習MILAAIAPT
- 綜述論文：當前深度神經網路模型壓縮和加速方法速覽神經網路模型
- Rank & Sort Loss for Object Detection and Instance Segmentation 論文解讀（含核心原始碼詳解）ObjectSegmentation原始碼
- 遷移學習《Cluster-Guided Semi-Supervised Domain Adaptation for Imbalanced Medical Image Classification》遷移學習GUIIDEAIAPT