Increasing Trust in Data-Driven Model Validation

Jöckel, Lisa; Kläs, Michael

doi:10.1007/978-3-030-26601-1_11

Cited by 12 publications

(6 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to change an image to augment a certain quality deficit, we have to understand how the deficit manifests in an image [5]. Two different approaches are illustrated in the following.…”

Section: A Quality Deficit Manifestationmentioning

confidence: 99%

“…b) Influence of Scope Characteristics: Besides intensity, characteristics of the specific application situation can also influence other aspects of the appearance of a quality deficit. Wind, e.g., causes a slant depending on the direction of the driving and the wind, and the type of environment (e.g., forest or city) affects the color of the dirt on a sign [5]. c) Influence of Quality Deficit Combination: The presence of one quality deficit can change the appearance of another.…”

Section: B Parameter Determinationmentioning

confidence: 99%

“…When it is dark, steam on a lens will also appear darker than otherwise. Thus, an application order of quality deficit augmentations and the effects between multiple quality deficits present have to be considered [5].…”

Section: B Parameter Determinationmentioning

confidence: 99%

“…In traffic sign recognition, possible quality deficits caused by context conditions, the sign itself, or the camera sensor were identified in [5]. Examples are weather conditions, physical damage, dirt on the lens, or lighting conditions.…”

Section: Introductionmentioning

confidence: 99%

“…In this context, we propose an approach for creating methods for augmentation that can be applied in combination, and whose components can be adapted and reused for further quality deficits. The proposed approach can be utilized to provide augmentation functionality in frameworks like [5].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Safe Traffic Sign Recognition through Data Augmentation for Autonomous Vehicles Software

Jöckel

Kläs

Martínez‐Fernández

2019

2019 IEEE 19th International Conference on Software Quality, Reliability and Security Companion (QRS-C)

View full text Add to dashboard Cite

Since autonomous vehicles operate in an open context, their software components, including data-driven ones, have to reliably process inputs (e.g., obtained by cameras) in order to make safe decisions. A key challenge when providing reliable data-driven components is insufficient training data, which could lead to wrong interpretation of the environment, thereby causing accidents. Aim: The goal of our research is to extend available training data of data-driven components for safe autonomous vehicles using the example of traffic sign recognition. Method: We developed an approach to create realistic image augmentations of various quality deficits and applied them on the German traffic sign recognition benchmark dataset (GTSRB). Results: The approach results in images augmented with (any combination of) seven different quality deficits affecting traffic sign recognition (rain, dirt on lens, steam on lens, darkness, motion blur, dirt on sign, backlight) and considers dependencies between combined quality deficits and influences from other contextual information. Conclusion: Our approach can be used to obtain more comprehensive datasets, especially also including samples with quality deficits that are difficult to gather. By structuring the augmentation into a set of basic components, the approach can be adapted for other application domains (e.g., person detection).

show abstract

Section: A Quality Deficit Manifestationmentioning

confidence: 99%

Section: B Parameter Determinationmentioning

confidence: 99%

Section: B Parameter Determinationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations