Domain Adaptation for Semantic Segmentation of Urban Scenes

By Yang Zhang, Philip David, Boqing Gong. 2017

Subfields: Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Representation & Self-Supervised Learning

Citations in the corpus

Domain Adaptation for Traffic Density Estimation (2021, methodology): “…Domain Shift - TRANCOS Dataset (Guerrero-Gómez-Olmedo et al., 2015) Hydra-CNN (Oñoro-Rubio and López-Sastre, 2016) 10.99 68.70 0.71 FCN-MT (Zhang et al., 2017a) 5.31 - 0.85 LC-ResFCN (Laradji et al., 2018) 3.32 - - Baseline - CSRNet (Li et al., 2018) 3.56 30.64 0.10 Our Approach 3.30…”
Unsupervised Model Adaptation for Continual Semantic Segmentation (2020, methodology): “Distributional alignment can be achieved by matching the distributions at different levels of abstraction, including appearance (Hoffman et al. 2018a; Sankaranarayanan et al. 2018), feature (Hoffman et al. 2018a; Murez et al. 2018), output (Zhang, David, and Gong 2017) levels.”
Source Data-Absent Unsupervised Domain Adaptation Through Hypothesis Transfer and Labeling Transfer (2020, unknown): “nition [2], [4], [5], [6], semantic segmentation [3], [7], [8], [9], sentiment classification [10], [11], object detection [12], [13], and person re-identification [14], [15].”
A Fully Convolutional Tri-Branch Network (FCTN) for Domain Adaptation (2017, background): “al [5] proposed to predict the class distribution over the entire image and some representative super pixels in the target domain first.”
FDA: Fourier Domain Adaptation for Semantic Segmentation (2020, background): “[52] proposes curriculumstyle learning to align both global distributions over images”
Unsupervised vehicle counting via multiple camera domain adaptation (2020, background): “However, as pointed out in [14], they do not perform well in other tasks, such as semantic segmentation.”
Conditional Generative Adversarial Network for Structured Domain Adaptation (2018, methodology): “The other four work [21, 45, 9, 8] focus on domain adaptation to enhance the segmentation performance on real images by networks trained on virtual ones, which is also our concern in this work.”
Multi-Modal Domain Adaptation for Fine-Grained Action Recognition (2020, background): “UDA has been extensively studied for vision tasks including object recognition [14, 15, 32, 33, 51, 55], semantic segmentation [18, 60, 65] and person reidentification [10, 49, 62].”
A Survey of Unsupervised Deep Domain Adaptation (2018, background): “[276] develop a curriculum where after learning some high-level properties on easy tasks the segmentation network is forced to follow those properties (interpretations include student-teacher setup or posterior regularization), and Perone et al.”
Gradually Vanishing Bridge for Adversarial Domain Adaptation (2020, background): “In traditional UDA methods, domain discrepancy is reduced by learning transferable representations [10, 14, 28], or reweighing the importance of the samples [29].”
Curriculum Learning: A Survey (2021, methodology): “The third largest homogeneous cluster (red) is mostly composed of domain adaptation methods [67], [84], [45], [109], [130], [136], [28], [155], [145], [115], [153].”
CrDoCo: Pixel-Level Domain Transfer With Cross-Domain Consistency (2019, result): “We compare our approach with the state-of-theart methods [41, 51, 24, 15, 14, 7, 17, 35, 47].”
ROAD: Reality Oriented Adaptation for Semantic Segmentation of Urban Scenes (2017, methodology): “9% mean IoU, which is similar to those reported in [17, 49], where VGG-16 was also used as the backbone network.”
SplitNet: Sim2Sim and Task2Task Transfer for Embodied Visual Navigation (2019, background): “Sim2Real: Significant progress has been made on adapting between simulated and real imagery for visual recognition, especially in the context of semantic segmentation in driving scenes [10, 11, 12, 39].”
Learning from Synthetic Data: Addressing Domain Shift for Semantic Segmentation (2017, methodology): “Both approaches have had good success in the classification problems; however, as pointed out in [32], their performance improvement does not translate well to the semantic segmentation problem.”
UM-Adapt: Unsupervised Multi-Task Adaptation Using Adversarial Cross-Task Distillation (2019, background): “Although such approaches are very effective in classification or recognition based tasks [56], they yield suboptimal performance for adaptation of fully-convolutional architectures, which is particularly essential for spatial prediction tasks [62].”
Domain Adaptation for Structured Output via Discriminative Patch Representations (2019, methodology): “, VGG-16 and ResNet-101, and compare with state-of-the-art approaches via feature adaptation [16, 46], pixel-level adaptation [15], pseudo label re-training [50] and output space alignment [5, 39].”