-
CAD Models to Real-World Images: A Practical Approach to Unsupervised Domain Adaptation in Industrial Object Classification
Authors:
Dennis Ritter,
Mike Hemberger,
Marc Hönig,
Volker Stopp,
Erik Rodner,
Kristian Hildebrand
Abstract:
In this paper, we systematically analyze unsupervised domain adaptation pipelines for object classification in a challenging industrial setting. In contrast to standard natural object benchmarks existing in the field, our results highlight the most important design choices when only category-labeled CAD models are available but classification needs to be done with real-world images. Our domain ada…
▽ More
In this paper, we systematically analyze unsupervised domain adaptation pipelines for object classification in a challenging industrial setting. In contrast to standard natural object benchmarks existing in the field, our results highlight the most important design choices when only category-labeled CAD models are available but classification needs to be done with real-world images. Our domain adaptation pipeline achieves SoTA performance on the VisDA benchmark, but more importantly, drastically improves recognition performance on our new open industrial dataset comprised of 102 mechanical parts. We conclude with a set of guidelines that are relevant for practitioners needing to apply state-of-the-art unsupervised domain adaptation in practice. Our code is available at https://github.com/dritter-bht/synthnet-transfer-learning.
△ Less
Submitted 7 October, 2023;
originally announced October 2023.
-
Particle Identification by Multifractal Parameters in Gamma-Astronomy with the HEGRA-Cherenkov-Telescopes
Authors:
B. M. Schaefer,
W. Hofmann,
H. Lampeitl,
M. Hemberger
Abstract:
Cherenkov images of air showers can also be classified using multifractal and wavelet parameters, as compared to the conventional Hillas image parameters. This technique was applied to the images recorded by the cameras of the stereoscopic imaging air Cherenkov-telescopes operated by the HEGRA collaboration. With respect to the identification of particles, the performance of multifractal and wav…
▽ More
Cherenkov images of air showers can also be classified using multifractal and wavelet parameters, as compared to the conventional Hillas image parameters. This technique was applied to the images recorded by the cameras of the stereoscopic imaging air Cherenkov-telescopes operated by the HEGRA collaboration. With respect to the identification of particles, the performance of multifractal and wavelet parameters was examined using a data sample from the observation of the active galaxy Mkn 501 that showed a high gamma-ray flux. The multifractal parameters were also combined with the Hillas parameters using a neural network approach in order to further improve the gamma/hadron-separation.
△ Less
Submitted 18 January, 2001;
originally announced January 2001.
-
Effectiveness of TeV Gamma-Ray Observations at Large Zenith Angles with a Stereoscopic System of Imaging Atmospheric Cherenkov Telescopes
Authors:
A. Konopelko,
F. Aharonian,
M. Hemberger,
W. Hofmann,
J. Kettler,
G. Puehlhofer,
H. J. Voelk
Abstract:
The sensitivity of imaging atmospheric Cherenkov telescopes (IACTs) in TeV gamma-ray observations reachs its maximum at small zenith angles (< 30 degree) which provide the minimum attainable energy threshold of an instrument. However, for a specific telescope site a number of gamma-ray sources, or source candidates, can only be observed at much larger zenith angles (< 60 degree). Moreover the ob…
▽ More
The sensitivity of imaging atmospheric Cherenkov telescopes (IACTs) in TeV gamma-ray observations reachs its maximum at small zenith angles (< 30 degree) which provide the minimum attainable energy threshold of an instrument. However, for a specific telescope site a number of gamma-ray sources, or source candidates, can only be observed at much larger zenith angles (< 60 degree). Moreover the observations at large zenith angles allow to extend the observation time window for any object seen at small zenith angles, as well as to enlarge the dynamic energy range of an instrument towards the highest observable energies of gamma-rays. Based on Monte Carlo simulations we present here the results on the sensitivity of a stereoscopic system of 5 IACTs in observations at large zenith angles. We point out some important parameters of the telescope design which could substantially improve the efficiency of such observations with forthcoming IACT arrays like CANGAROO III, HESS and VERITAS.
△ Less
Submitted 25 June, 1999;
originally announced June 1999.
-
Performance of the Stereoscopic System of the HEGRA Imaging Air Cherenkov Telescopes: Monte Carlo Simulations & Observations
Authors:
A. Konopelko,
M. Hemberger,
F. Aharonian
Abstract:
Based on the Monte Carlo simulations we have studied the performance of the HEGRA system of imaging air Cherenkov telescopes (IACTs) in its present configuration of 4 IACTs as well as in its future final configuration of 5 IACTs. Here we present the results on the basic characteristics of the IACT system which are used in the standard data analysis procedure, i.e., the collection areas, the dete…
▽ More
Based on the Monte Carlo simulations we have studied the performance of the HEGRA system of imaging air Cherenkov telescopes (IACTs) in its present configuration of 4 IACTs as well as in its future final configuration of 5 IACTs. Here we present the results on the basic characteristics of the IACT system which are used in the standard data analysis procedure, i.e., the collection areas, the detection rates, the angular resolution, the energy resolution, and the gamma/hadron-separation efficiency. By comparing several key Monte Carlo predictions with experimental results it is possible to check the accuracy of the simulations. The Monte Carlo results concerning hadron-nuclear showers are tested with the recorded cosmic ray events and the results concerning photon-induced showers are tested with a large data sample of gamma-rays observed from BL Lac object Mkn 501 during its high flaring activity in 1997. Summarizing the simulations and current observations we give the basic recommendations of using the instrument and the major values of its sensitivity.
△ Less
Submitted 15 January, 1999;
originally announced January 1999.
