Convolutional Neural Networks

Publikation: Beiträge in SammelwerkenKapitel

Authors

Dieses Kapitel führt in Convolutional Neural Networks (CNNs) ein und beschreibt, wie diese im Kontext der Sportanalyse verwendet werden können. Insbesondere eignen sich CNNs für das End-to-End-Lernen auf Bildern oder ähnlich strukturierten Daten. Dabei können CNNs Merkmale von Bildern anhand der Pixelwerte effizient lernen und beispielsweise sehr gute Merkmale für eine Klassifikationsaufgabe extrahieren. Die Modelle profitieren dabei von Parameter Sharing in den Convolutional Layern und zeichnen sich durch (bedingte) Translationsäquivarianz und -invarianz aus. CNNs eignen sich auch dafür, Merkmale aus Positionsdaten von Teamsportarten zu lernen, sofern die Daten in eine entsprechende Struktur gebracht werden.
OriginalspracheEnglisch
TitelSportinformatik : Modellbildung, Simulation, Datenanalyse und Visualisierung von sportbezogenen Daten
HerausgeberDaniel Memmert
Anzahl der Seiten9
ErscheinungsortBerlin
VerlagSpringer Spektrum
Erscheinungsdatum01.01.2023
Seiten207-215
ISBN (Print)978-3-662-67025-5
ISBN (elektronisch)978-3-662-67026-2
DOIs
PublikationsstatusErschienen - 01.01.2023

DOI

Zuletzt angesehen

Publikationen

  1. Guest Editorial - ''Econometrics of Anonymized Micro Data''
  2. Development of a Parameterized Model for Additively Manufactured Dies to Control the Strains in Extrudates
  3. Formative Perspectives on the Relation Between CSR Communication and CSR Practices
  4. Robust approximate fixed-time tracking control for uncertain robot manipulators
  5. Performance Saga: Interview 01
  6. A Sliding Mode Control with a Bang-Bang Observer for Detection of Particle Pollution
  7. Learning from partially annotated sequences
  8. Other spaces
  9. Does thinking-aloud affect learning, visual information processing and cognitive load when learning with seductive details as expected from self-regulation perspective?
  10. Kalman Filter for Predictive Maintenance and Anomaly Detection
  11. Applications of the Simultaneous Modular Approach in the Field of Material Flow Analysis
  12. Active learning for network intrusion detection
  13. Complexity of traffic scenes and EEG-measures of processing workload in car driving
  14. Material flow during constrained friction processing and its effects on the local properties of AM50 rods
  15. Mirrored piezo servo hydraulic actuators for use in camless combustion engines and its Control with mirrored inputs and MPC
  16. Scale-dependent diversity patterns affect spider assemblages of two contrasting forest ecosystems
  17. An Outcome-Oriented, Social-Ecological Framework for Assessing Protected Area Effectiveness
  18. The Creation of the Concept through the Interaction of Philosophy with Science and Art
  19. From "cracking the orthographic code" to "playing with language"
  20. Project and Design of a Catamaran Prototype with Aerial Propulsion System
  21. Intraspecific trait variation increases species diversity in a trait-based grassland model
  22. Cost effectiveness of guided Internet-based interventions for depression in comparison with control conditions
  23. Studying properties of water data using manifold-aware anomaly detectors
  24. Constraints are the solution, not the problem
  25. Orchestrating distributed data governance in open social innovation
  26. Teaching methods for modelling problems and students’ task-specific enjoyment, value, interest and self-efficacy expectations
  27. Self-regulation in error management training: emotion control and metacognition as mediators of performance effects
  28. Understanding and Supporting Management Decision-Making
  29. Topic selection and development in learner-native speaker voice-based telecollaborative discourse
  30. Effect of gap distortion on the field splitting of collective modes in superfluid He3-B
  31. Estimation and interpretation of a Heckman selection model with endogenous covariates