Testing of Image Processing Algorithms on Synthetic Data

Neumann-Cosel, Kilian von; Roth, Erwin; Lehmann, Daniel; Speth, Johannes; Knoll, Alois

doi:10.1109/icsea.2009.34

Cited by 7 publications

(3 citation statements)

References 2 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There is little literature in which the presence and/or repercussions of geographic representation bias in RSI datasets is either mentioned [54][55][56] or investigated [33,34]. Descriptions of new RSI datasets often reference the need for model's likely performance on real-world data, including any limits of model reliability due to spatial effects [121][122][123][124][125][126][127].…”

Section: A Current Challengesmentioning

confidence: 99%

Methods for Measuring Geodiversity in Large Overhead Imagery Datasets

Wesley

Matisziw

2021

IEEE Access

View full text Add to dashboard Cite

Geographic variation in the appearance of objects on Earth is readily observable in remotely sensed imagery (RSI) and somewhat intuitive to understand for most peoplemany classes of objects (houses, vehicles, crop fields etc.) simply look different depending on their location. This variation has recently been shown to have important implications when training machine learning models on geotagged image datasets for specific object detection and classification tasks. For example, models trained on datasets with ethnocentric biases in image content have been shown to misclassify objects in under-sampled regions, particularly in least-developed countries. The need to evaluate the growing corpus of RSI datasets for representativeness, heterogeneity and geodiversity is therefore high; yet scalable methods for measuring these concepts are absent in the remote sensing domain. This paper introduces the first dataset analysis methods for detecting and assessing geodiversity problems in large RSI datasets, based on geospatial adaptations of the Fréchet Inception Distance and Inception Score in the deep learning framework. Geospatial Fréchet Distance is proposed as a dissimilarity measure for image features of an object class across geographic regions useful for comparing differences in object class appearance in different locations and/or spatial scales. A complementary Geospatial Inception Score is proposed to quantify heterogeneity of geographic context present in dataset labels within particular regions/locations, taking into account the labels themselves as well as their immediate surroundings. Rigorous tests of these methods on simulated RSI datasets demonstrate their stability, sensitivity, and the broad range of dataset analyses to which they can be applied.INDEX TERMS Big geodata, deep convolutional neural networks, explainable ai, geographic domain shift, GIS-remote sensing fusion, imagery interpretation, spatial data analysis This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.

show abstract

Section: A Current Challengesmentioning

confidence: 99%

Methods for Measuring Geodiversity in Large Overhead Imagery Datasets

Wesley

Matisziw

2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Here, we discuss some of the previous works related to vision problems. In [22], Neumann-Cosel et al use synthetic data obtained from the modeled Virtual Test Drive simulation for lane tracking in driver assistance and active safety systems. An identical work from Haltakov et al [8] proposes a framework for generation of synthetic data from a realistic driving simulator to create modeled traffic scenarios similar to real-world scenarios.…”

Section: Related Workmentioning

confidence: 99%

Low-cost scene modeling using a density function improves segmentation performance

Sharma

Yayilgan

Gool

2016

2016 25th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN)

View full text Add to dashboard Cite

We propose a low cost and effective way to combine a free simulation software and free CAD models for modeling human-object interaction in order to improve human & object segmentation. It is intended for research scenarios related to safe human-robot collaboration (SHRC) and interaction (SHRI) in the industrial domain. The task of human and object modeling has been used for detecting activity, and for inferring and predicting actions, different from those works, we do human and object modeling in order to learn interactions in RGB-D data for improving segmentation. For this purpose, we define a novel density function to model a three dimensional (3D) scene in a virtual environment (VREP). This density function takes into account various possible configurations of human-object and object-object relationships and interactions governed by their affordances. Using this function, we synthesize a large, realistic and highly varied synthetic RGB-D dataset that we use for training. We train a random forest classifier, and the pixelwise predictions obtained is integrated as a unary term in a pairwise conditional random fields (CRF). Our evaluation shows that modeling these interactions improves segmentation performance by ∼7% in mean average precision and recall over state-of-the-art methods that ignore these interactions in realworld data. Our approach is computationally efficient, robust and can run real-time on consumer hardware.

show abstract

“…In order to test different types of ADAS, Dirndorfer, v. Neumann-Cosel et al also designed a simulation framework, whose main simulation components and experiences, i.e. "Virtual Test Drive" (VTD) [19] and "Automotive Data-and Time-triggered Framework" (ADTF) were transferred to the here outlined approach [4,12,13]. The framework differs from Berger's aforementioned who designed an environment entirely from scratch.…”

Section: Related Workmentioning

confidence: 99%

Simulations on Consumer Tests

Block¹,

Heeren²,

Kühnel³

et al. 2013

Proceedings of International Workshop on Engineering Simulations for Cyber-Physical Systems

View full text Add to dashboard Cite

This article discusses new challenges for series development regarding the vehicle safety that arise from the recently published AEB test protocol by the consumer-test-organisation EuroNCAP for driver assistance systems [6]. The tests from the test protocol are of great significance for an OEM that sells millions of cars each year, due to the fact that a positive rating of the vehicle-under-test (VUT) in safety relevant aspects is important for the reputation of a car manufacturer. The further intensification and aggravation of the test requirements for those systems is one of the challenges, that has to be mastered in order to continuously make significant contributions to safety for high-volume cars. Therefore, it is to be shown how a simulation approach may support the development process, especially with tolerance analysis. This article discusses the current stage of work, steps that are planned for the future and results that can be expected at the end of such an analysis.

show abstract

Testing of Image Processing Algorithms on Synthetic Data

Cited by 7 publications

References 2 publications

Methods for Measuring Geodiversity in Large Overhead Imagery Datasets

Methods for Measuring Geodiversity in Large Overhead Imagery Datasets

Low-cost scene modeling using a density function improves segmentation performance

Simulations on Consumer Tests

Contact Info

Product

Resources

About