Deteksi Objek Berbahaya pada Video Animasi Kartun Anak Menggunakan YOLO26 Berbasis Computer Vision
DOI:
https://doi.org/10.24252/instek.v11i2.69650Keywords:
Animasi Kartun, YOLO26, Deteksi Objek, Objek Berbahaya, Computer VisionAbstract
Perkembangan media digital meningkatkan akses anak terhadap tayangan animasi kartun yang tidak selalu aman, di mana beberapa tayangan diketahui mengandung objek berbahaya seperti pisau, senjata tajam, pistol, dan senapan yang berpotensi memberikan pengaruh negatif terhadap perkembangan anak. Penelitian ini bertujuan mengimplementasikan sistem deteksi objek berbahaya pada video animasi kartun anak menggunakan algoritma YOLO26 (You Only Look Once) untuk filtrasi konten. Kontribusi utama penelitian ini adalah pembangunan dataset anotasi khusus berisi 4.486 gambar dengan 5.060 bounding box yang dikumpulkan dari berbagai gaya animasi kartun, berbeda dari penelitian terdahulu yang umumnya menggunakan dataset domain nyata (real-world). Dataset terdiri atas empat kelas objek (pisau, senjata tajam, pistol, dan senapan) dan dibagi dengan rasio 80:10:10 untuk data latih, validasi, dan uji, dengan model YOLO26 di-fine-tuning dari bobot pra-latih COCO. Model YOLO26 dilatih dan dievaluasi menggunakan metrik Precision, Recall, F1-Score, [email protected], dan [email protected]:0.95 terhadap 449 gambar uji dengan 501 objek berbahaya. Hasil evaluasi menunjukkan Precision sebesar 0,945, Recall 0,938, F1-Score 0,941, [email protected] sebesar 0,953, dan [email protected]:0.95 sebesar 0,763, dengan [email protected] di atas 0,88 pada seluruh kelas. Model diimplementasikan sebagai purwarupa aplikasi berbasis web untuk filtrasi konten yang memungkinkan pengguna memasukkan tautan video animasi dari repositori video aplikasi dan memperoleh hasil deteksi otomatis.
Downloads
References
[1] P. T. Giraffe Insights, "PARK: Precise Advertiser Report - Kids.", Giraffe Insights & Precise TV, America (USA), Market Research Report, Jun 2022. [Daring]. Tersedia pada: https://content.precisify.com/hubfs/PARK_%20Precise%20Advertiser%20Report%20-%20Kids%20%20.pdf
[2] M. B. R. Victoria Rideout, M.A. Ph.D., "The Common Sense census: Media use by kids age zero to eight, 2020," Common Sense Media, San Francisco, CA, 2020. [Daring]. Tersedia pada: https://www.commonsensemedia.org/sites/default/files/research/report/2020_zero_to_eight_census_final_web.pdf
[3] M. Gkolemi, P. Papadopoulos, E. P. Markatos, dan N. Kourtellis, "YouTubers Not madeForKids: Detecting Channels Sharing Inappropriate Videos Targeting Children," dalam 14th ACM Web Science Conference 2022, Jun 2022, hlm. 370-381. doi: 10.1145/3501247.3531556.
[4] Beverly Hills, California, America. Tom and Jerry | Mega Compilation | Vol. 20 | Warner Classics | Sufferin' Cats, (1943). [Daring Video]. Tersedia pada: https://www.youtube.com/watch?v=LNAEv9t5HWI
[5] "Peraturan Komisi Penyiaran Indonesia Nomor 03 Tahun 2007 tentang Standar Program Siaran," Komisi Penyiaran Indonesia, Jakarta, 2007. [Daring]. Tersedia pada: https://www.kpi.go.id/download/regulasi/Standar%20Program%20Siaran.pdf
[6] X. Zhao, L. Wang, Y. Zhang, X. Han, M. Deveci, dan M. Parmar, "A review of convolutional neural networks in computer vision," Artif. Intell. Rev., vol. 57, no. 4, hlm. 99, Mar 2024, doi: 10.1007/s10462-024-10721-6.
[7] R. Sapkota dkk., "YOLO advances to its genesis: a decadal and comprehensive review of the You Only Look Once (YOLO) series," Artif. Intell. Rev., vol. 58, no. 9, hlm. 274, Jun 2025, doi: 10.1007/s10462-025-11253-3.
[8] W. Xu, D. Zhu, R. Deng, K. Yung, dan A. W. H. Ip, "Violence-YOLO: Enhanced GELAN Algorithm for Violence Detection," Appl. Sci., vol. 14, no. 15, hlm. 6712, Agu 2024, doi: 10.3390/app14156712.
[9] Y. Zeng, L. Zhang, J. Zhao, J. Lan, dan B. Li, "JRL-YOLO: A Novel Jump-Join Repetitious Learning Structure for Real-Time Dangerous Object Detection," Comput. Intell. Neurosci., vol. 2021, no. 1, hlm. 5536152, Jan 2021, doi: 10.1155/2021/5536152.
[10] G. Lavanya dan S. D. Pande, "Enhancing Real-time Object Detection with YOLO Algorithm," EAI Endorsed Trans. Internet Things, vol. 10, Des 2023, doi: 10.4108/eetiot.4541.
[11] M. A. R. Alif dan M. Hussain, "YOLOv1 to YOLOv10: A comprehensive review of YOLO variants and their application in the agricultural domain," 14 Juni 2024, arXiv: arXiv:2406.10139. doi: 10.48550/arXiv.2406.10139.
[12] L. Jiao dkk., "A Survey of Deep Learning-Based Object Detection," IEEE Access, vol. 7, hlm. 128837-128868, 2019, doi: 10.1109/ACCESS.2019.2939201.
[13] X. Mao dkk., "COCO-O: A Benchmark for Object Detectors under Natural Distribution Shifts," 2 Agustus 2023, arXiv: arXiv:2307.12730. doi: 10.48550/arXiv.2307.12730.
[14] D. Kadish, S. Risi, dan A. S. Lovlie, "Improving Object Detection in Art Images Using Only Style Transfer," 4 Mei 2021, arXiv: arXiv:2102.06529. doi: 10.48550/arXiv.2102.06529.
[15] Y. Tang, "AnimeNet: A Deep Learning Approach for Detecting Violence and Eroticism in Animated Content," Comput. Mater. Contin., vol. 77, no. 1, hlm. 867-891, 2023, doi: 10.32604/cmc.2023.041550.
[16] T. Khalil, S. Jan, R. M. Ghoniem, dan M. I. K. Khalil, "Multi-CNN Fusion Framework for Predictive Violence Detection in Animated Media," Comput. Mater. Contin., vol. 86, no. 2, hlm. 1-20, 2026, doi: 10.32604/cmc.2025.072655.
[17] M. Taha, A. Alsammak, dan A. Zaky, "InspectorNet: Transformer network for violence detection in animated cartoon," Eng. Res. J. - Fac. Eng. Shoubra, vol. 52, no. 2, hlm. 114-119, Apr 2023, doi: 10.21608/erjsh.2023.181713.1119.
[18] F. Zhuang dkk., "A Comprehensive Survey on Transfer Learning," 23 Juni 2020, arXiv: arXiv:1911.02685. doi: 10.48550/arXiv.1911.02685.
[19] J. Yosinski, J. Clune, Y. Bengio, dan H. Lipson, "How transferable are features in deep neural networks?," 6 November 2014, arXiv: arXiv:1411.1792. doi: 10.48550/arXiv.1411.1792.
[20] K. Yousaf dan T. Nawaz, "A Deep Learning-Based Approach for Inappropriate Content Detection and Classification of YouTube Videos," IEEE Access, vol. 10, hlm. 16283-16298, 2022, doi: 10.1109/ACCESS.2022.3147519.
[21] K. Abhinand dkk., "Malayalam Sign Language Identification using Finetuned YOLOv8 and Computer Vision Techniques," 8 Mei 2024, arXiv: arXiv:2405.06702. doi: 10.48550/arXiv.2405.06702.
[22] P. Hidayatullah dan R. Tubagus, "YOLO26: A Comprehensive Architecture Overview and Key Improvements," 2026, doi: https://doi.org/10.48550/arXiv.2602.14582.
[23] S. Ren, K. He, R. Girshick, dan J. Sun, "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks," 6 Januari 2016, arXiv: arXiv:1506.01497. doi: 10.48550/arXiv.1506.01497.
[24] Amresh Kumar, "Fine-tuning YOLO for Custom Object Detection: A Complete Guide," Medium. [Daring]. Tersedia pada: https://medium.com/@amresh.kumar11/fine-tuning-yolo-for-custom-object-detection-a-complete-guide-6ce3724ce9f1
[25] Aston Zhang, Zachary C. Lipton, dan A. J. S. Mu Li, "Dive into Deep Learning," hlm. 1185, 2023, doi: arXiv:2106.11342.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Ridwan Widiyono, Muslim Hidayat, Iman Ahmad Ihsannuddin
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
