Evaluation of Traffic Sign Recognition Methods Trained on Synthetically Generated Data

Moiseev, B. N.; Konev, Artem; Chigorin, Alexander; Konushin, Anton

doi:10.1007/978-3-319-02895-8_52

Cited by 66 publications

(37 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this section, we evaluate the performance of the proposed SLC-AE on seven benchmark data sets, i.e., COIL20 [47], MNIST, USPS, SYN Signs [48], GTSRB [49], VOC 2007 [50] and MSRC [51]. Detailed descriptions of the seven data sets and experimental setting are presented in Section IV-A; the comparative results are reported in Section IV-B; and the in-depth analysis, including the parameter studies and ablation studies, is presented in Section IV-C and Section IV-D respectively.…”

Section: Methodsmentioning

confidence: 99%

Domain Adaptation by Stacked Local Constraint Auto-Encoder Learning

Peng

Murphey

et al. 2019

IEEE Access

View full text Add to dashboard Cite

Domain adaptation (DA), a particular case of transfer learning, is an effective technology for learning a discriminative model in scenarios where the data from the training (source) and the testing (target) domains share common class labels but follow different distributions. The differences between domains, called domain shifts, are caused by variations in the acquisition devices and environmental conditions, such as changing illuminations, pose, and collecting-device noises, that are related to a specific domain, denoted as domain-specific noises in this paper. The research on stacked denoising autoencoder (SDA) has demonstrated that noise-robust features could be learned through training a model to reduce the manmade (simulated) noises. However, little research has been conducted to learn the domain-invariant features through training SDA to reduce the domain-specific noises from the real word. In this paper, we propose a novel variant of SDA for DA, called the stacked local constraint auto-encoder (SLC-AE), which aims to learn domain-invariant features through iteratively optimizing the SDA and the low-dimensional manifold. The core idea behind the SLC-AE is that both the source and target samples are corrupted due to the domain-specific noises, and each corrupted sample could be de-noised by calculating the weighted sum of its neighbor samples defined on the intrinsic manifold. Because the neighbor samples on the intrinsic manifold are semantically similar, their weighted sum preserves the generic information and reduces the domain-specific noises. To properly evaluate the performance of the SLC-AE, we conducted extensive experiments using seven benchmark data sets, i.e., MNIST, USPS, COIL20, SYN SIGNS, GTSRB, MSRC and VOC 2007. Compared to twelve different state-of-the-art methods, the experimental results demonstrated that the proposed SLC-AE model made significant improvement over the performance of SDA and achieved the best average performance on the seven data sets.

show abstract

Section: Methodsmentioning

confidence: 99%

Domain Adaptation by Stacked Local Constraint Auto-Encoder Learning

Peng

Murphey

et al. 2019

IEEE Access

View full text Add to dashboard Cite

show abstract

“…• Synthetic Signs [38]: this dataset was synthetically generated by taking common street signs from Wikipedia and applying several transformations. It tries to simulate images from GTSRB although there are significant differences between them.…”

Section: Datasetsmentioning

confidence: 99%

Incremental Unsupervised Domain-Adversarial Training of Neural Networks

Gallego

Calvo-Zaragoza

Fisher

2021

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

In the context of supervised statistical learning, it is typically assumed that the training set comes from the same distribution that draws the test samples. When this is not the case, the behavior of the learned model is unpredictable and becomes dependent upon the degree of similarity between the distribution of the training set and the distribution of the test set. One of the research topics that investigates this scenario is referred to as Domain Adaptation (DA). Deep neural networks brought dramatic advances in pattern recognition and that is why there have been many attempts to provide good domain adaptation algorithms for these models. Here we take a different avenue and approach the problem from an incremental point of view, where the model is adapted to the new domain iteratively. We make use of an existing unsupervised domain-adaptation algorithm to identify the target samples on which there is greater confidence about their true label. The output of the model is analyzed in different ways to determine the candidate samples. The selected samples are then added to the source training set by self-labeling, and the process is repeated until all target samples are labeled. This approach implements a form of adversarial training in which, by moving the self-labeled samples from the target to the source set, the DA algorithm is forced to look for new features after each iteration. Our results report a clear improvement with respect to the nonincremental case in several datasets, also outperforming other state-of-the-art domain adaptation algorithms.

show abstract

“…In this task we recognize traffic signs from german traffic sign recognition benchmark (GTSRB) [56] by adapting from labeled synthesized images [40]. In total, 90K labeled source images, 35K unlabeled target images for training, 430 labeled target images for validation, 12, 569 labeled target images for testing are used.…”

Section: Synthetic Signs→gtsrbmentioning

confidence: 99%

“…For all tasks, we apply the same data preprocessing of channel-wise mean and standard deviation normalization per example [18], i.e.,x Table S11: Evaluation on digit and traffic sign adaptation tasks, such as MNIST [29] to MNIST-M [13] (M→MM), Synthetic Digits [13] to SVHN (S→S), SVHN to MNIST (S→M), MNIST to SVHN (M→S), or Synthetic Signs [40] to GTSRB [56] (S→G). Experiments are executed for 10 times with different random seeds and mean test set accuracy and standard error are reported.…”

Section: Synthetic Signs→gtsrbmentioning

confidence: 99%

Gotta Adapt 'Em All: Joint Pixel and Feature-Level Domain Adaptation for Recognition in the Wild

Tran

Sohn

et al. 2019

2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)

View full text Add to dashboard Cite

Recent developments in deep domain adaptation have allowed knowledge transfer from a labeled source domain to an unlabeled target domain at the level of intermediate features or input pixels. We propose that advantages may be derived by combining them, in the form of different insights that lead to a novel design and complementary properties that result in better performance. At the feature level, inspired by insights from semi-supervised learning, we propose a classification-aware domain adversarial neural network that brings target examples into more classifiable regions of source domain. Next, we posit that computer vision insights are more amenable to injection at the pixel level. In particular, we use 3D geometry and image synthesis based on a generalized appearance flow to preserve identity across pose transformations, while using an attribute-conditioned Cycle-GAN to translate a single source into multiple target images that differ in lower-level properties such as lighting. Besides standard UDA benchmark, we validate on a novel and apt problem of car recognition in unlabeled surveillance images using labeled images from the web, handling explicitly specified, nameable factors of variation through pixel-level and implicit, unspecified factors through feature-level adaptation.

show abstract

Evaluation of Traffic Sign Recognition Methods Trained on Synthetically Generated Data

Cited by 66 publications

References 18 publications

Domain Adaptation by Stacked Local Constraint Auto-Encoder Learning

Domain Adaptation by Stacked Local Constraint Auto-Encoder Learning

Incremental Unsupervised Domain-Adversarial Training of Neural Networks

Gotta Adapt 'Em All: Joint Pixel and Feature-Level Domain Adaptation for Recognition in the Wild

Contact Info

Product

Resources

About