Semi-supervised semantic segmentation needs strong, high-dimensional perturbations

By Geoff French, Timo Aila, Samuli Laine, Michal Mackiewicz, Graham David Finlayson. 2019

137 citations recorded by the source corpus.

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

ZeroWaste Dataset: Towards Deformable Object Segmentation in Cluttered Scenes (2021, background): “limitations by jointly learning from both annotated and unannotated images [11, 22, 29, 40, 42, 45].”
Revisiting Weak-to-Strong Consistency in Semi-Supervised Semantic Segmentation (2022, methodology): “Following closely the research line of semi-supervised learning (SSL), advanced methods in semisupervised semantic segmentation have evolved from GANs-based adversarial training paradigm [21, 56, 47] into the widely adopted consistency regularization framework [19, 49, 81, 12, 28, 29, 62] and reborn”
Semi-Supervised and Unsupervised Deep Visual Learning: A Survey (2022, methodology): “…generally inherit similar learning rationales as SSL or UL for image classiﬁcation, and adapt techniques such as consistency regularization [31], [178], [179], [180], self-training [32], [168], [174], [181], [182], [183], [184], GAN frameworks [185], [186] in SSL, or contrastive learning [187],…”
ST++: Make Self-trainingWork Better for Semi-supervised Semantic Segmentation (2021, methodology): “It is worth noting that, our method with ResNet-50 backbone even surpasses [16, 58] with ResNet-101 backbone by a large margin.”
DACS: Domain Adaptation via Cross-domain Mixed Sampling (2020, methodology): “ annotate some data from the domain of interest and re-train (ﬁne-tune) the network on this new data. Additionally, Semi-Supervised Learning (SSL) methods for semantic segmentation have been proposed [4, 5, 6, 7, 8, 9], effectively training on a small amount of labelled data by relying on complement”
PseudoSeg: Designing Pseudo Labels for Semantic Segmentation (2020, background): “, augmentation) for image segmentation to enable consistency training framework (French et al., 2020; Ouali et al., 2020), we show that a wise design of pseudo labels for segmentation has great veiled potentials.”
Bootstrapping Semantic Segmentation with Regional Contrast (2021, methodology): “…strong semi-supervised segmentation methods ourselves: S4GAN (Mittal et al., 2019): an adversarial learning based semi-supervised method; CutOut (French et al., 2020), CutMix (French et al., 2020), ClassMix (Olsson et al., 2021): three image augmentation strategies designed speciﬁcally for…”
Semi-Supervised Semantic Segmentation with Cross Pseudo Supervision (2021, methodology): “In addition, we study the single-network pseudo supervision in a way like [11] with the CutMix augmentation.”
ClassMix: Segmentation-Based Data Augmentation for Semi-Supervised Learning (2020, methodology): “Recent works have addressed this issue by either applying perturbations on an encoded state of the network instead of the input [7], or by using the augmentation technique CutMix [8] to enforce consistent predictions over mixed samples [6, 9].”
Milking CowMask for Semi-Supervised Image Classification (2020, methodology): “We follow French et al. (2019) when we use the term CowOut to refer to CowMask used for erasure and CowMix when used for mixing.”