Towards Corner Case Detection by Modeling the Uncertainty of Instance Segmentation Networks

Heidecker, Florian; Hannan, Abdul; Bieshaar, Maarten; Sick, Bernhard

doi:10.1007/978-3-030-68799-1_26

Cited by 11 publications

(6 citation statements)

References 26 publications

(37 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As an example for a method level corner case, consider a ML method for visual object detection. While we might aim to detect unknown objects by a high epistemic uncertainty [17], also typical "normal" objects can spark a high epistemic uncertainty [14], [18], hence, leading to a method level corner case. Epistemic uncertainty is not necessarily limited to ML methods and can appear in various kinds of mathematical models in general.…”

Section: Methods Layer Corner Casesmentioning

confidence: 99%

An Application-Driven Conceptualization of Corner Cases for Perception in Highly Automated Driving

Heidecker,

Breitenstein,

Rösch

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Systems and functions that rely on machine learning (ML) are the basis of highly automated driving. An essential task of such ML models is to reliably detect and interpret unusual, new, and potentially dangerous situations. The detection of those situations, which we refer to as corner cases, is highly relevant for successfully developing, applying, and validating automotive perception functions in future vehicles where multiple sensor modalities will be used. A complication for the development of corner case detectors is the lack of consistent definitions, terms, and corner case descriptions, especially when taking into account various automotive sensors. In this work, we provide an application-driven view of corner cases in highly automated driving. To achieve this goal, we first consider existing definitions from the general outlier, novelty, anomaly, and out-of-distribution detection to show relations and differences to corner cases. Moreover, we extend an existing camera-focused systematization of corner cases by adding RADAR (radio detection and ranging) and LiDAR (light detection and ranging) sensors. For this, we describe an exemplary toolchain for data acquisition and processing, highlighting the interfaces of the corner case detection. We also define a novel level of corner cases, the method layer corner cases, which appear due to uncertainty inherent in the methodology or the data distribution.

show abstract

Section: Methods Layer Corner Casesmentioning

confidence: 99%

An Application-Driven Conceptualization of Corner Cases for Perception in Highly Automated Driving

Heidecker,

Breitenstein,

Rösch

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Among others, methods like Monte Carlo dropout [9,79], deep ensemble [12,80], Bayes-by-Backprop [10], or Prior networks [11] aim to approximate the epistemic uncertainty. A high epistemic uncertainty of detected objects, classes, or other predictions is considered as an indicator for corner cases [78,81,82] because the ML model cannot reliably handle these data samples. In this respect, we have to consider the approximation quality of the predicted uncertainties [83].…”

Section: Model Levelmentioning

confidence: 99%

Corner cases in machine learning processes

Heidecker,

Bieshaar,

Sick

2024

AI Perspect. Adv.

Self Cite

View full text Add to dashboard Cite

Applications using machine learning (ML), such as highly autonomous driving, depend highly on the performance of the ML model. The data amount and quality used for model training and validation are crucial. If the model cannot detect and interpret a new, rare, or perhaps dangerous situation, often referred to as a corner case, we will likely blame the data for not being good enough or too small in number. However, the implemented ML model and its associated architecture also influence the behavior. Therefore, the occurrence of prediction errors resulting from the ML model itself is not surprising. This work addresses a corner case definition from an ML model’s perspective to determine which aspects must be considered. To achieve this goal, we present an overview of properties for corner cases that are beneficial for the description, explanation, reproduction, or synthetic generation of corner cases. To define ML corner cases, we review different considerations in the literature and summarize them in a general description and mathematical formulation, whereby the expected relevance-weighted loss is the key to distinguishing corner cases from common data. Moreover, we show how to operationalize the corner case characteristics to determine the value of a corner case. To conclude, we present the extended taxonomy for ML corner cases by adding the input, model, and deployment levels, considering the influence of the corner case properties.

show abstract

“…This definition has been refined into a systematization of corner cases in camera data [2], and all available sensors in automated vehicles [3]. This refinement allows for specific development of methods treating certain corner case types, e.g., detecting an anomalous amount of objects in a scene [4], or unknown objects [5], [6].…”

Section: Introductionmentioning

confidence: 99%

Amodal Cityscapes: A New Dataset, its Generation, and an Amodal Semantic Segmentation Challenge Baseline

Breitenstein¹,

Fingscheidt²

2022

Preprint

View full text Add to dashboard Cite

Amodal perception terms the ability of humans to imagine the entire shapes of occluded objects. This gives humans an advantage to keep track of everything that is going on, especially in crowded situations. Typical perception functions, however, lack amodal perception abilities and are therefore at a disadvantage in situations with occlusions. Complex urban driving scenarios often experience many different types of occlusions and, therefore, amodal perception for automated vehicles is an important task to investigate. In this paper, we consider the task of amodal semantic segmentation and propose a generic way to generate datasets to train amodal semantic segmentation methods. We use this approach to generate an amodal Cityscapes dataset. Moreover, we propose and evaluate a method as baseline on Amodal Cityscapes, showing its applicability for amodal semantic segmentation in automotive environment perception. We provide the means to re-generate this dataset on github 1 .

show abstract

Towards Corner Case Detection by Modeling the Uncertainty of Instance Segmentation Networks

Cited by 11 publications

References 26 publications

An Application-Driven Conceptualization of Corner Cases for Perception in Highly Automated Driving

An Application-Driven Conceptualization of Corner Cases for Perception in Highly Automated Driving

Corner cases in machine learning processes

Amodal Cityscapes: A New Dataset, its Generation, and an Amodal Semantic Segmentation Challenge Baseline

Contact Info

Product

Resources

About