WebAnnotate-Motion+: Anotação Temporal Assistida e Orientada à Privacidade de Atividades Humanas em Vídeos naWeb

• Julio C. S. Lima UFLA
• Daniel Henrique Silva UFLA
• Esther Silva de Magalhães UFLA
• Luiz H. A. Correia UFLA
• André de Lima Salgado UFLA
• Mayron C. O. Moreira UFLA

DOI: https://doi.org/10.5753/webmedia_estendido.2026.34747

Resumo

Temporal annotation of human activities in videos is essential for multimedia datasets and human activity recognition (HAR), but manual labeling can be labor-intensive and may involve sensitive recordings. This paper presents WebAnnotate-Motion+, a privacy-oriented Web tool for human-reviewed temporal annotation. Video decoding and analysis take place in the browser; the server stores project metadata, temporal segments, review states, and analysis records rather than the audiovisual file. Two local mechanisms assist the annotator: visual frame variation and body-motion estimation from pose landmarks. Their outputs remain separate from validated annotations and can be accepted, rejected, or adjusted by the user. The tool also supports overlapping intervals, project persistence, and JSON/CSV export. Functional validation exercised the integrated workflow, but no controlled user study was conducted; therefore, productivity improvements are not claimed. WebAnnotate-Motion+ offers an extensible basis for investigating privacy-conscious, human-reviewed temporal annotation.

Palavras-chave: video annotation, human activity recognition, multimedia authoring, privacy, human-in-the-loop, Web application

Referências

Valentin Bazarevsky, Ivan Grishchenko, Karthik Raveendran, Tyler Zhu, Fan Zhang, and Matthias Grundmann. 2020. BlazePose: On-device Real-time Body Pose Tracking. arXiv preprint arXiv:2006.10204 (2020). DOI: 10.48550/arXiv.2006.10204

Bruna Carolina Rodrigues Cunha, Olibário J. Machado Neto, and Maria da Graça Campos Pimentel. 2013. A Heuristic Evaluation of a Mobile Annotation Tool. In Proceedings of the 19th Brazilian Symposium on Multimedia and theWeb (WebMedia). Sociedade Brasileira de Computação, Porto Alegre, Brazil, 89–92.

CVAT.ai Corporation. 2026. Computer Vision Annotation Tool (CVAT). DOI: 10.5281/zenodo.21494920

Abhishek Dutta and Andrew Zisserman. 2019. The VIA Annotation Software for Images, Audio and Video. In Proceedings of the 27th ACM International Conference on Multimedia. Association for Computing Machinery, New York, NY, USA, 2276–2279. DOI: 10.1145/3343031.3350535

Fabian Caba Heilbron and Juan Carlos Niebles. 2014. Collecting and Annotating Human Activities in Web Videos. In Proceedings of the International Conference on Multimedia Retrieval. Association for Computing Machinery, New York, NY, USA, 377–384. DOI: 10.1145/2578726.2578775

Camillo Lugaresi, Jiuqiang Tang, Hadon Nash, Chris McClanahan, Esha Uboweja, Michael Hays, Fan Zhang, Chuo-Ling Chang, Ming Guang Yong, Juhyun Lee,Wan-Teh Chang, Wei Hua, Manfred Georg, and Matthias Grundmann. 2019. MediaPipe: A Framework for Building Perception Pipelines. arXiv preprint arXiv:1906.08172 (2019). DOI: 10.48550/arXiv.1906.08172

Massachusetts Institute of Technology. 1988. MIT License. [link]. Acesso em: 11 ago. 2026.

Hao Quan and Andrea Bonarini. 2022. HAVPTAT: A Human Activity Video Pose Tracking Annotation Tool. Software Impacts 12 (2022), 100278. DOI: 10.1016/j.simpa.2022.100278

Hang Zhao, Antonio Torralba, Lorenzo Torresani, and Zhicheng Yan. 2019. HACS: Human Action Clips and Segments Dataset for Recognition and Temporal Localization. In Proceedings of the IEEE/CVF International Conference on Computer Vision. 8668–8678. DOI: 10.1109/ICCV.2019.00876