Computer Vision for MOBA Analytics: A Dataset and Baseline for Visibility Analysis in Dota 2

• Ricardo da Rocha Carvalho PUCPR
• Eloísa Oliveira PUCPR
• Luiz Bernardo Martins Kummer PUCPR
• Emerson Cabrera Paraiso PUCPR
• Rayson Laroca PUCPR

DOI: https://doi.org/10.5753/sbgames.2026.26163

Resumo

Introduction: Most Multiplayer Online Battle Arena (MOBA) analytics studies rely on structured data, which does not directly capture what each team could actually see during a match. Objective: This work introduces Dota2-Vis, a video-based dataset, and a baseline pipeline for visibility analysis in professional Dota 2 matches. Methodology: The dataset comprises all 144 matches from The International 2025, recorded from both team perspectives, totaling 288 Full HD videos, together with 2,477 manually annotated minimap images. We evaluate multiple variants of a modern object detector for player-icon detection and use the best-performing model to estimate opponent-visible player presence over time. Results: YOLO11l (large) achieved the best overall performance, reliably identifying player icons even in dense and visually cluttered minimap scenes. The resulting visibility curves reveal player, hero, role, and team-level patterns that complement conventional MOBA analytics, highlighting behavioral differences that are difficult to obtain from structured data alone. The Dota2-Vis dataset is publicly available at https://github.com/RicardoRCarvalho/dota2-vis/.

Palavras-chave: Computer Vision, Dota 2, Esports, Game Analytics

Referências

Barz, B. e Denzler, J. (2020). Do we train on test data? Purging CIFAR of near-duplicates. Journal of Imaging, 6(6):41.

Białecki, A., Michalak, B., e Gajewski, J. (2024). Esports training, periodization, and software—a scoping review. Applied Sciences, 14(22):10354.

Chitayat, A. P., Kokkinakis, A., Patra, S., Demediuk, S., Robertson, J., Olarewaju, O., Ursu, M., Kirmann, B., Hook, J., Block, F., e Drachen, A. (2020). WARDS: Modelling the worth of vision in MOBA’s. In Science and Information Conference, pages 63–81.

Costa, L. M., Drachen, A., Souza, F. C. M., e Xexéo, G. (2024). Artificial intelligence in MOBA games: A multivocal literature mapping. IEEE Transactions on Games, 16(2):250–269.

Everingham, M., Van Gool, L., Williams, C. K. I., Winn, J., e Zisserman, A. (2010). The pascal visual object classes (VOC) challenge. International Journal of Computer Vision, 88(2):303–338.

Hodge, V. J., Devlin, S., Sephton, N., Block, F., Cowling, P. I., e Drachen, A. (2021). Win prediction in multiplayer esports: Live professional match prediction. IEEE Transactions on Games, 13(4):368–379.

Joo, H.-T., Lee, S.-H., Chung, I., e Kim, K.-J. (2025). Improving esports viewing experience through hierarchical scene detection and tracking. Scientific Reports, 15(1):9803.

Katona, A., Spick, R., Hodge, V. J., Demediuk, S., Block, F., Drachen, A., e Walker, J. A. (2019). Time to die: Death prediction in Dota 2 using deep learning. In IEEE Conference on Games (CoG), pages 1–8.

Kim, D.-W., Park, S.-Y., Yang, S.-I., e Lee, S.-K. (2025). Real-time player tracking framework on MOBA game video through object detection. IEEE Transactions on Games, 17(2):498–509.

Laroca, R., dos Santos, M., e Menotti, D. (2025). Improving small drone detection through multi-scale processing and data augmentation. In International Joint Conference on Neural Networks (IJCNN), pages 1–8.

Li, Q., Xu, P., Chan, Y. Y., Wang, Y., Wang, Z., Qu, H., e Ma, X. (2017). A visual analytics approach for understanding reasons behind snowballing and comeback in MOBA games. IEEE Transactions on Visualization and Computer Graphics, 23(1):211–220.

Lin, T.-Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., Dollár, P., e Zitnick, C. L. (2014). Microsoft COCO: Common objects in context. In European Conference on Computer Vision (ECCV), pages 740–755.

Liquipedia (2026). The International. [link]. Accessed: 2026-04-18.

Novak, A. R., Bennett, K. J. M., Pluss, M. A., e Fransen, J. (2019). Performance analysis in esports: Part 1 – the validity and reliability of match statistics and notational analysis in League of Legends. SportRxiv.

Ringer, C., Missaoui, S., Hodge, V. J., Chitayat, A. P., Kokkinakis, A., Patra, S., Demediuk, S., Munoz, A. C., Olarewaju, O., Ursu, M., Kirman, B., Hook, J., Block, F., Drachen, A., e Walker, J. A. (2023). Time to die 2: Improved in-game death prediction in Dota 2. Machine Learning with Applications, 12:100466.

Seif El-Nasr, M., Drachen, A., e Canossa, A., editors (2013). Game Analytics: Maximizing the Value of Player Data. Springer, London.

Semenov, A., Romov, P., Korolev, S., Yashkov, D., e Neklyudov, K. (2017). Performance of machine learning algorithms in predicting game outcome from drafts in Dota 2. In International Conf. on Analysis of Images, Social Networks and Texts, pages 26–37.

Summerville, A., Cook, M., e Steenhuisen, B. (2016). Draft-analysis of the ancients: predicting draft picks in Dota 2 using machine learning. In AAAI Conference on Artificial Intelligence and Interactive Digital Entertainment, pages 100–106.

Terven, J., Córdova-Esparza, D.-M., e Romero-González, J.-A. (2023). A comprehensive review of YOLO architectures in computer vision: From YOLOv1 to YOLOv8 and YOLO-NAS. Machine Learning and Knowledge Extraction, 5(4):1680–1716.

Tot, M., Conserva, M., Chitayat, A. P., Kokkinakis, A., Patra, S., Demediuk, S., Munoz, A. C., Olarewaju, O., Ursu, M., Kirmann, B., Hook, J., Block, F., Drachen, A., e Perez-Liebana, D. (2021). What are you looking at? Team fight prediction through player camera. In IEEE Conference on Games (CoG), pages 1–8.

Ultralytics (2024). YOLOv11. [link]. Accessed: 2026-04-26.

Yang, Z., Pan, Z., Wang, Y., Cai, D., Shi, S., Huang, S.-L., Bi, W., e Liu, X. (2022). Interpretable real-time win prediction for Honor of Kings—a popular mobile MOBA esport. IEEE Transactions on Games, 14(4):589–597.

Yang, Z., Wang, Y., Li, P., Lin, S., Shi, S., Huang, S.-L., e Bi, W. (2023). Predicting events in MOBA games: Prediction, attribution, and evaluation. IEEE Transactions on Games, 15(2):193–201.