Advanced Pedestrian Dataset Augmentation for Autonomous Driving

Vobecký, Antonín; Uřičář, Michal; Hurych, David; Škoviera, Radoslav

doi:10.1109/iccvw.2019.00290

Cited by 8 publications

(6 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Based on the prior knowledge of camera parameters, Hattori et al [70] proposed to generate a variety of geometrically accurate images of synthetic pedestrians. Vobecky et al [185] used the technique of GAN [61] to generate people images in a required pose according to specific pose key-points. Wu et al [203] developed a multi-modal cascaded generative adversarial network with U-net structure to generate pedestrian data.…”

Section: Pure Cnn Based Pedestrian Detection Methodsmentioning

confidence: 99%

From Handcrafted to Deep Features for Pedestrian Detection: A Survey

Cao

Pang

Xie

et al. 2022

IEEE Trans. Pattern Anal. Mach. Intell.

View full text Add to dashboard Cite

Pedestrian detection is an important but challenging problem in computer vision, especially in human-centric tasks. Over the past decade, significant improvement has been witnessed with the help of handcrafted features and deep features. Here we present a comprehensive survey on recent advances in pedestrian detection. First, we provide a detailed review of single-spectral pedestrian detection that includes handcrafted features based methods and deep features based approaches. For handcrafted features based methods, we present an extensive review of approaches and find that handcrafted features with large freedom degrees in shape and space have better performance. In the case of deep features based approaches, we split them into pure CNN based methods and those employing both handcrafted and CNN based features. We give the statistical analysis and tendency of these methods, where feature enhanced, part-aware, and post-processing methods have attracted main attention. In addition to single-spectral pedestrian detection, we also review multi-spectral pedestrian detection, which provides more robust features for illumination variance. Furthermore, we introduce some related datasets and evaluation metrics, and compare some representative methods. We conclude this survey by emphasizing open problems that need to be addressed and highlighting various future directions. Researchers can track an up-to-date list at https://github.com/JialeCao001/PedSurvey.

show abstract

Section: Pure Cnn Based Pedestrian Detection Methodsmentioning

confidence: 99%

From Handcrafted to Deep Features for Pedestrian Detection: A Survey

Cao

Pang

Xie

et al. 2022

IEEE Trans. Pattern Anal. Mach. Intell.

View full text Add to dashboard Cite

show abstract

“…To mitigate such an effect, several works propose to integrate constraints to counteract this unwanted behavior [12]. Some works utilize generative networks for augmentation [21,30,28] to extend the variation of pedestrian instances. In practice, various GAN variants [31,12,14] are proven capable of providing a set of samples whose distribution is similar to the target data.…”

Section: Related Workmentioning

confidence: 99%

Attention-based Adversarial Appearance Learning of Augmented Pedestrians

Strauss,

Savkin,

Tombari

2021

Preprint

View full text Add to dashboard Cite

Synthetic data became already an essential component of machine learning-based perception in the field of autonomous driving. Yet it still cannot replace real data completely due to the sim2real domain shift. In this work, we propose a method that leverages the advantages of the augmentation process and adversarial training to synthesize realistic data for the pedestrian recognition task. Our approach utilizes an attention mechanism driven by an adversarial loss to learn domain discrepancies and improve sim2real adaptation. Our experiments confirm that the proposed adaptation method is robust to such discrepancies and reveals both visual realism and semantic consistency. Furthermore, we evaluate our data generation pipeline on the task of pedestrian recognition and demonstrate that generated data resemble properties of the real domain.

show abstract

“…However, there is a possibility of offering false classifications, predictions, and denial of services based on the AI model's data processing and decision-making ability. The perceptive impact of AI-based decision-making has been observed in specific services and use cases for certain population groups [1], [27], [34]. Examples of gender and racial biases in healthcare, banking, and the hiring process have been discussed by providing potential mitigating strategies [27].…”

Section: Introductionmentioning

confidence: 99%

“…In another use case, self-driving car object detection algorithms failed to predict specific user groups as the datasets used for training the AI model consisted of class representation from humans with white race [1]. Similarly, self-driving vehicles may also show biased results during the classification and detection of women and mobility-impaired individuals, as the datasets, especially the validation set, lack such representation of classes [34]. These observations have led to the inclusion of Biases in the decision-making process are mainly inherited in AI models because of close-world assumption (CWA), which is generally used for knowledge representation in the datasets [34].…”

Section: Introductionmentioning

confidence: 99%

“…Similarly, self-driving vehicles may also show biased results during the classification and detection of women and mobility-impaired individuals, as the datasets, especially the validation set, lack such representation of classes [34]. These observations have led to the inclusion of Biases in the decision-making process are mainly inherited in AI models because of close-world assumption (CWA), which is generally used for knowledge representation in the datasets [34]. In CWA, the test/validation set samples are assumed to be similar to the training set.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Bias Detection and Generalization in AI Algorithms on Edge for Autonomous Driving

Katare

Kourtellis

Park

et al. 2022

2022 IEEE/ACM 7th Symposium on Edge Computing (SEC)

View full text Add to dashboard Cite

A machine learning model can often produce biased outputs for a familiar group or similar sets of classes during inference over an unknown dataset. The generalization of neural networks have been studied to resolve biases, which has also shown improvement in accuracy and performance metrics, such as precision and recall, and refining the dataset's validation set. Data distribution and instances included in test and validation-set play a significant role in improving the generalization of neural networks. For producing an unbiased AI model, it should not only be trained to achieve high accuracy and minimize false positives. The goal should be to prevent the dominance of one class/feature over the other class/feature while calculating weights. This paper investigates state-of-art object detection/classification on AI models using metrics such as selectivity score and cosine similarity. We focus on perception tasks for vehicular edge scenarios, which generally include collaborative tasks and model updates based on weights. The analysis is performed using cases that include the difference in data diversity, the viewpoint of the input class and combinations. Our results show the potential of using cosine similarity, selectivity score and invariance for measuring the training bias, which sheds light on developing unbiased AI models for future vehicular edge services.

show abstract

Advanced Pedestrian Dataset Augmentation for Autonomous Driving

Cited by 8 publications

References 9 publications

From Handcrafted to Deep Features for Pedestrian Detection: A Survey

From Handcrafted to Deep Features for Pedestrian Detection: A Survey

Attention-based Adversarial Appearance Learning of Augmented Pedestrians

Bias Detection and Generalization in AI Algorithms on Edge for Autonomous Driving

Contact Info

Product

Resources

About