PENINGKATAN RESOLUSI VIDEO MUSIK LAWAS INDONESIA MENGGUNAKAN REAL ESRGAN X4PLUS
DOI:
https://doi.org/10.51401/jinteks.v7i2.6146Keywords:
Real-ESRGAN, Super-Resolusi, Restorasi Video, PSNR, SSIMAbstract
Video musik Indonesia lawas yang diproduksi pada era VCD umumnya memiliki kualitas visual rendah dan degradasi artefak yang signifikan. Penelitian ini mengevaluasi efektivitas model Real?ESRGAN?X4plus dalam meningkatkan resolusi video-video tersebut menggunakan pendekatan super?resolusi berbasis pembelajaran mendalam. Eksperimen dilakukan pada lima video dengan resolusi awal 480p yang ditingkatkan menjadi 1920p, dengan variasi parameter denoise_strength dari 0,0 hingga 1,0 dan tile_size dari 64 hingga 256. Model ini dibandingkan dengan metode interpolasi bicubic dan ESRGAN publik sebagai baseline. Evaluasi dilakukan menggunakan metrik PSNR dan SSIM, serta uji statistik paired t?test. Hasil menunjukkan bahwa konfigurasi optimal adalah denoise_strength = 0.0 dan tile_size = 256, menghasilkan PSNR tertinggi (35,10 dB) dan SSIM terbaik (0,956). Dibanding baseline, Real?ESRGAN memberikan peningkatan kualitas yang signifikan baik secara obyektif maupun visual. Model ini juga menunjukkan potensi dalam mempertahankan detail halus dan konsistensi temporal antar frame. Temuan ini mendukung pemanfaatan Real?ESRGAN sebagai solusi restorasi video budaya Indonesia, serta membuka peluang untuk pengembangan lanjut berbasis VSR temporal.
References
M. Casad, O. Rieger, and D. Alexander, “Enduring access to rich media content: Understanding use and usability requirements,” D?Lib Mag., vol. 21, no. 9/10, 2015, doi: 10.1045/september2015?casad.
A. Das, D. Paramaj, and S. BR, “Scalable video fidelity enhancement: Leveraging the state?of?the?art AI models,” Scalable Comput. Pract. Exp., vol. 25, no. 3, pp. 1658–1666, 2024, doi: 10.12694/scpe.v25i3.2696.
H. Han and J. Lv, “Super?resolution?empowered adaptive medical video streaming in telemedicine systems,” Electronics, vol. 11, no. 18, Art. no. 2944, 2022, doi: 10.3390/electronics11182944.
T. Higaki et al., “Deep learning reconstruction at CT: Phantom study of the image characteristics,” Acad. Radiol., vol. 27, no. 1, pp. 82–87, 2020, doi: 10.1016/j.acra.2019.09.008.
M. Hussain, S. Abbas, and M. Irshad, “Quadratic trigonometric B?spline for image interpolation using GA,” PLoS ONE, vol. 12, no. 6, Art. no. e0179721, 2017, doi: 10.1371/journal.pone.0179721.
S. Kim, H. Jang, J. Jang, Y. Lee, and D. Hwang, “Deep?learned short tau inversion recovery imaging using multi?contrast MR images,” Magn. Reson. Med., vol. 84, no. 6, pp. 2994–3008, 2020, doi: 10.1002/mrm.28327.
H. Liu, P. Zhao, Z. Ruan, F. Shang, and Y. Liu, “Large motion video super?resolution with dual subnet and multi?stage communicated upsampling,” in Proc. AAAI Conf. Artif. Intell., vol. 35, no. 3, pp. 2127–2135, 2021, doi: 10.1609/aaai.v35i3.16310.
D. Madhu, S. Vasuhi, and A. Samydurai, “Dynamic 8?bit XOR algorithm with AES crypto algorithm for image steganography,” Signal Image Video Process., vol. 18, Suppl. 1, pp. 429–445, 2024, doi: 10.1007/s11760?024?03165?6.
H. Martínez, A. Hines, and M. Farias, “Perceptual quality of audio?visual content with common video and audio degradations,” Appl. Sci., vol. 11, no. 13, Art. no. 5813, 2021, doi: 10.3390/app11135813.
Z. Salah, W. Al?Sit, Z. Salah, and E. Elsoud, “Spatial domain noise removal filtering for low?resolution digital images,” Indones. J. Electr. Eng. Comput. Sci., vol. 34, no. 3, pp. 1627–1642, 2024, doi: 10.11591/ijeecs.v34.i3.pp1627?1642.
T. Shanableh, “Recursive quad?tree block partitioning for data embedding in images,” SN Comput. Sci., vol. 1, no. 6, Art. no. 457, 2020, doi: 10.1007/s42979?020?00322?9.
A. Shukla et al., “A novel zero?shot real world spatio?temporal super?resolution (ZS?RW?TSR) model for video super?resolution,” IEEE Access, vol. 12, pp. 123969–123984, 2024, doi: 10.1109/ACCESS.2024.3406476.
X. Tang, F. Ouyang, Y. Xu, L. Zhu, and B. Peng, “A unified model fusing region of interest detection and super resolution for video compression,” Comput. Mater. Continua, vol. 79, no. 3, pp. 3955–3975, 2024, doi: 10.32604/cmc.2024.049057.
S. Wang, A. Politis, A. Mesaros, and T. Virtanen, “Self?supervised learning of audio representations from audio?visual data using spatial alignment,” IEEE J. Sel. Topics Signal Process., vol. 16, no. 6, pp. 1467–1479, 2022, doi: 10.1109/JSTSP.2022.3180592.
X. Wang, L. Xie, C. Dong, and Y. Shan, “Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data,” arXiv preprint arXiv:2107.10833, Jul. 2021. [Online]. Available: https://arxiv.org/abs/2107.10833
F. Winarni, “Aspek hukum peran serta masyarakat dalam pelestarian cagar budaya,” Mimbar Hukum, vol. 30, no. 1, pp. 94–104, 2018, doi: 10.22146/jmh.29160.
Y. Xu, T. Guo, and C. Wang, “A remote sensing image super?resolution reconstruction model combining multiple attention mechanisms,” Sensors, vol. 24, no. 14, Art. no. 4492, 2024, doi: 10.3390/s24144492.
H. Yao et al., “DR2?Net: Deep residual reconstruction network for image compressive sensing,” Neurocomputing, vol. 359, pp. 483–493, 2019, doi: 10.1016/j.neucom.2019.05.006.
M. Zerva and L. Kondi, “Video super?resolution using plug?and?play priors,” IEEE Access, vol. 12, pp. 11963–11971, 2024, doi: 10.1109/ACCESS.2024.3355195.
Published
How to Cite
Issue
Section
Copyright (c) 2025 Abdul Rauf Alfansani, Ema Utami, Dhani Ariatmanto
This work is licensed under a Creative Commons Attribution 4.0 International License.
