Generalize then Adapt: Source-Free Domain Adaptive Semantic Segmentation

Kundu, Jogendra Nath; Kulkarni, Akshay; Singh, Amit Kumar; Jampani, Varun; Babu, R. Venkatesh

doi:10.1109/iccv48922.2021.00696

Cited by 68 publications

(27 citation statements)

References 75 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This method, also described by Algorithm 1, is simply executed as one forward pass for each new target domain image x D T t . Note that in contrast to previous source-free UDA methods [54], [55], [56] our CBNA method is a continual source-free UDA method (cf. Fig.…”

Section: Novel Continuous Batchnorm Adaptation (Cbna)mentioning

confidence: 97%

Continual BatchNorm Adaptation (CBNA) for Semantic Segmentation

Klingner¹,

Ayache²,

Fingscheidt³

2022

Preprint

View full text Add to dashboard Cite

Environment perception in autonomous driving vehicles often heavily relies on deep neural networks (DNNs), which are subject to domain shifts, leading to a significantly decreased performance during DNN deployment. Usually, this problem is addressed by unsupervised domain adaptation (UDA) approaches trained either simultaneously on source and target domain datasets or even source-free only on target data in an offline fashion. In this work, we further expand a source-free UDA approach to a continual and therefore online-capable UDA on a single-image basis for semantic segmentation. Accordingly, our method only requires the pre-trained model from the supplier (trained in the source domain) and the current (unlabeled target domain) camera image. Our method Continual BatchNorm Adaptation (CBNA) modifies the source domain statistics in the batch normalization layers, using target domain images in an unsupervised fashion, which yields consistent performance improvements during inference. Thereby, in contrast to existing works, our approach can be applied to improve a DNN continuously on a single-image basis during deployment without access to source data, without algorithmic delay, and nearly without computational overhead. We show the consistent effectiveness of our method across a wide variety of source/target domain settings for semantic segmentation. As part of this work, our code will be made publicly available. 1

show abstract

Section: Novel Continuous Batchnorm Adaptation (Cbna)mentioning

confidence: 97%

Continual BatchNorm Adaptation (CBNA) for Semantic Segmentation

Klingner¹,

Ayache²,

Fingscheidt³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…We denote the problem of learning multiple labeled target domains without retaining previously seen data as domain-incremental learning (DIL) [50], [51], [52], [53], [54], [55], [56], [57]. Another sub-category called source-free UDA -sometimes referred to as unsupervised model adaptation (UMA) -has recently gained increasing attention for both classification [58], [59], [60], [61] and segmentation tasks [62], [63], [64], [65], [66]. These methods perform UDA from a source to a single target domain without using the actual source data.…”

Section: Sequential Udamentioning

confidence: 99%

Generative appearance replay for continual unsupervised domain adaptation

Chen¹,

Thandiackal²,

Pati³

et al. 2023

Preprint

View full text Add to dashboard Cite

Deep learning models can achieve high accuracy when trained on large amounts of labeled data. However, real-world scenarios often involve several challenges: Training data may become available in installments, may originate from multiple different domains, and may not contain labels for training. Certain settings, for instance medical applications, often involve further restrictions that prohibit retention of previously seen data due to privacy regulations. In this work, to address such challenges, we study unsupervised segmentation in continual learning scenarios that involve domain shift. To that end, we introduce GarDA (Generative Appearance Replay for continual Domain Adaptation), a generative-replay based approach that can adapt a segmentation model sequentially to new domains with unlabeled data. In contrast to single-step unsupervised domain adaptation (UDA), continual adaptation to a sequence of domains enables leveraging and consolidation of information from multiple domains. Unlike previous approaches in incremental UDA, our method does not require access to previously seen data, making it applicable in many practical scenarios. We evaluate GarDA on two datasets with different organs and modalities, where it substantially outperforms existing techniques.

show abstract

“…For this task some very recent methods have been developed concurrently: Teja et al [54] apply entropy minimization on the posterior and maximize the noise robustness of latent features. Kundu et al [55] use self-training on pseudo labels. Liu et al [56] also make use of this technique and in addition apply data-free knowledge distillation.…”

Section: Uda Without Source Datamentioning

confidence: 99%

Continual BatchNorm Adaptation (CBNA) for Semantic Segmentation

Klingner

Ayache

Fingscheidt

2022

IEEE Trans. Intell. Transport. Syst.

View full text Add to dashboard Cite

Environment perception in autonomous driving vehicles often heavily relies on deep neural networks (DNNs), which are subject to domain shifts, leading to a significantly decreased performance during DNN deployment. Usually, this problem is addressed by unsupervised domain adaptation (UDA) approaches trained either simultaneously on source and target domain datasets or even source-free only on target data in an offline fashion. In this work, we further expand a source-free UDA approach to a continual and therefore online-capable UDA on a single-image basis for semantic segmentation. Accordingly, our method only requires the pre-trained model from the supplier (trained in the source domain) and the current (unlabeled target domain) camera image. Our method Continual BatchNorm Adaptation (CBNA) modifies the source domain statistics in the batch normalization layers, using target domain images in an unsupervised fashion, which yields consistent performance improvements during inference. Thereby, in contrast to existing works, our approach can be applied to improve a DNN continuously on a single-image basis during deployment without access to source data, without algorithmic delay, and nearly without computational overhead. We show the consistent effectiveness of our method across a wide variety of source/target domain settings for semantic segmentation. Code is available at https://github.com/ifnspaml/CBNA

show abstract

Generalize then Adapt: Source-Free Domain Adaptive Semantic Segmentation

Cited by 68 publications

References 75 publications

Continual BatchNorm Adaptation (CBNA) for Semantic Segmentation

Continual BatchNorm Adaptation (CBNA) for Semantic Segmentation

Generative appearance replay for continual unsupervised domain adaptation

Continual BatchNorm Adaptation (CBNA) for Semantic Segmentation

Contact Info

Product

Resources

About