Real-Time Detection of Volleyball Player Movements with YOLOv8 to Support Rehabilitation

  • Maria Eduarda Caetano da Silva IFCE
  • William de Araújo Dias IFCE
  • Kaiuska Silva Peixoto UFC
  • Pedro Olavo de Paula Lima UFC
  • Antonio Augusto Teixeira Peixoto IFCE
  • Auzuir Ripardo de Alexandria IFCE

Resumo


This paper presents a computer vision system based on YOLOv8 for detecting and classifying critical volleyball postures, serve, set, pass (bump), and block, alongside ball tracking in real time. Designed to support physiotherapist and coaches, the system enables objective monitoring of biomechanical correctness during rehabilitation, addressing one of the key challenges in injury recovery: ensuring accurate execution of sport-specific movements. The project builds upon previous clinical research on volleyball rehabilitation and was implemented using open source tools, leveraging Google Colab for training on a custom-labeled dataset. Results demonstrate promising detection accuracy, indicating potential for integration into physiotherapy workflows and future augmented reality(AR) rehabilitation systems.
Palavras-chave: Computer Vision, YOLOv8, Volleyball, Physiotherapy, Posture Detection, Rehabilitation, Sports Analytics

Referências

PEIXOTO, Kaiuska Silva. Desenvolvimento e validação de um exergame com realidade virtual para reabilitação em praticantes de vôlei: um estudo piloto. 2025. Artigo.

GADE, Rikke; MERLER, Michele; THOMAS, Graham; MOESLUND, Thomas B. The (computer) vision of sports: Recent trends in research and commercial systems for sport analytics. Computer Vision, p. 296–311, 2024. Chapman and Hall/CRC.

REDMON, J.; FARHADI, A. YOLOv3: An Incremental Improvement. arXiv preprint arXiv:1804.02767, 2018.

JOCHER, G.; et al. Ultralytics YOLOv5 and YOLOv8. GitHub Repository, 2023. Disponível em: [link]. Acesso em: 11 ago. 2025.

SZELISKI, R. Computer Vision: Algorithms and Applications. Springer, 2022.

MATHIS, M.; et al. DeepLabCut: Markerless Pose Estimation. Nature Neuroscience, v. 21, p. 1281–1289, 2018.

ULTRALYTICS. YOLOv8: Modelo de IA da próxima geração da Ultralytics. 2023. Disponível em: [link]. Acesso em: 11 ago. 2025.

GOOGLE. Google Colab. 2025. Disponível em: [link]. Acesso em: 11 ago. 2025.

PYTHON SOFTWARE FOUNDATION. Python. 2025. Disponível em: [link]. Acesso em: 11 ago. 2025.

ROBOFLOW. Roboflow: Ferramentas de visão computacional para desenvolvedores e empresas. 2025. Disponível em: [link]. Acesso em: 11 ago. 2025.

ROBOFLOW. Dataset: dP3sBLA9Q4. 2025. Disponível em: [link]. Acesso em: 11 ago. 2025.

OPENCV. Open Source Computer Vision Library. 2025. Disponível em: [link]. Acesso em: 11 ago. 2025.

TCL. Tkinter: Python Interface to Tcl/Tk. 2025. Disponível em: [link]. Acesso em: 11 ago. 2025.

NVIDIA CORPORATION. NVIDIA. 2025. Disponível em: [link]. Acesso em: 11 ago. 2025.
Publicado
30/09/2025
SILVA, Maria Eduarda Caetano da; DIAS, William de Araújo; PEIXOTO, Kaiuska Silva; LIMA, Pedro Olavo de Paula; PEIXOTO, Antonio Augusto Teixeira; ALEXANDRIA, Auzuir Ripardo de. Real-Time Detection of Volleyball Player Movements with YOLOv8 to Support Rehabilitation. In: WORKSHOP DE TRABALHOS EM ANDAMENTO - CONFERENCE ON GRAPHICS, PATTERNS AND IMAGES (SIBGRAPI), 38. , 2025, Salvador/BA. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 2025 . p. 190-193.