Restoring historical murals for cultural heritage conservation remains difficult because these intricate artworks deteriorate severely throughout their existence. Preserving historical artwork requires precise techniques to both detect and digitally restore its deteriorated areas. The combination of manual restoration approaches and basic digital tools fails to deliver precise and efficient results, leading to inconsistent outcomes and the loss of original artistic details. This study develops an enhanced approach that integrates swin transformer for detecting damage with stable diffusion for performing the restoration process to achieve superior precision and reliability in restoration work. A dataset consisting of 1,200 high-resolution tamil nadu temple mural images which contains annotated degradation types for training and evaluating proposed models. Experimental results show that using swin transformer and stable diffusion produces superior results with an Intersection over Union (IoU) of 0.92 along with Peak Signal-to-Noise Ratio (PSNR) of 32.5 dB, Structural Similarity Index (SSIM) of 0.95 and Frechet Inception Distance (FID) of 12.3. The study validates the proposed methodology as an efficient system for enhancing the quality and mechanical fidelity in mural restoration applications which indicates its potential use as a solution for digital art restoration projects.

[1] Agarwal K. and Dixit M., “Scrupulous SCGAN Framework for Recognition of Restored Images with Caffe Based PCA Filtration,” The International Arab Journal of Information 352 The International Arab Journal of Information Technology, Vol. 23, No. 2, March 2026 Technology, vol. 21, no. 1, pp. 107-116, 2024. DOI: 10.34028/iajit/21/1/10

[2] Cai X., Lu Q., Yao J., Liu Y., and Hu Y., “An Ancient Murals Inpainting Method Based on Bidirectional Feature Adaptation and Adversarial Generative Networks,” Advances in Computer Graphics, vol. 14496, pp. 300-311, 2023. https://doi.org/10.1007/978-3-031-50072-5_24

[3] Cao J., Yan M., Jia Y., and et al., “Application of a Modified Inception-V3 Model in the Dynasty- Based Classification of Ancient Murals,” EURASIP Journal on Advances in Signal Processing, vol. 2021, no. 1, pp. 1-25, 2021. https://doi.org/10.1186/s13634-021-00740-8

[4] Choi Y., Murtala S., Jeong B., and Choi K., “Deep Learning-based Engraving Segmentation of 3-D Inscriptions Extracted from the Rough Surface of Ancient Stelae,” IEEE Access, vol. 9, pp. 153199- 153212, 2021. https://doi.org/10.1109/ACCESS.2021.3127229

[5] Gerasimiuk M., Shung D., Tong A., Stanley A., and et al., “MURAL: An Unsupervised Random Forest-based Embedding for Electronic Health Record Data,” in Proceedings of the International Conference on Big Data, Orlando, pp. 4694-4704, 2021. https://doi.org/10.1109/BigData52589.2021.9672 045

[6] Jiang X., Harun S., and Liu L., “Explainable Artificial Intelligence for Ancient Architecture and Lacquer Art,” Buildings, vol. 13, no. 5, pp. 1- 12, 2023. https://doi.org/10.3390/buildings13051213

[7] Lianji H. and Haihong Z., “Research and Application of Dunhuang Mural Image Restoration Technology Based on Deep Learning,” in Proceedings of the International Conference on Electronics and Devices, Computational Science, Marseille, pp. 348-353, 2024. https://doi.org/10.1109/ICEDCS64328.2024.00068

[8] Nasri A. and Huang X., “Images Enhancement of Ancient Mural Painting of Bey’s Palace Constantine, Algeria and Lacuna Extraction Using Mahalanobis Distance Classification Approach,” Sensors, vol. 22, no. 17, pp. 1-20, 2022. https://doi.org/10.3390/s22176643

[9] Pawar P., Ainapure B., Rashid M., Ahmad N., and et al., “Deep Learning Approach for the Detection of Noise Type in Ancient Images,” Sustainability, vol. 14, no. 18, pp. 1-19, 2022. https://doi.org/10.3390/su141811786

[10] Qian K., Zhang P., Huang R., Feng W., and Sun J., “Learning to Grade Deterioration for Ancient Murals,” Journal of Ambient Intelligence and Humanized Computing, vol. 15, pp. 1727-1734, 2024. https://doi.org/10.1007/s12652-019-01487-9

[11] Sharma R. and Kukreja V., “Detecting the Past: Advancements in Comic Panel Detection for Cultural Heritage Preservation,” in Proceedings of the 4th International Conference on Data Analytics for Business and Industry, Bahrain, pp. 529-532, 2023. https://doi.org/10.1109/ICDABI60145.2023.1062 9627

[12] Shi W. and Meng X., “Mural Restoration Research Based on Samples and Deep Learning,” in Proceedings of the IEEE International Conference on Information Technology, Electronics and Intelligent Communication Systems, Bangalore, pp. 1-4, 2024. https://doi.org/10.1109/ICITEICS61368.2024.10 625359

[13] Singh U., Maiti S., and Saini A., “Ancient Indian Murals Digital Restoration Through Image Inpainting,” in Proceedings of the 10th International Conference on Signal Processing and Integrated Networks, Noida, pp. 635-640, 2023. https://doi.org/10.1109/SPIN57001.2023.10116111

[14] Sun P., Hou M., Lyu S., Wang W., and et al., “Enhancement and Restoration of Scratched Murals Based on Hyperspectral Imaging-A Case Study of Murals in the Baoguang Hall of Qutan Temple, Qinghai, China,” Sensors, vol. 22, no. 24, pp. 1-19, 2022. https://doi.org/10.3390/s22249780

[15] Valdes J., Lopez C., and Lara A., “Detection of Deteriorated Sections in Murals Using Machine Learning,” in Proceedings of the IEEE 7th Congreso Internacional en Inteligencia Ambiental, Ingenieria de Software y Salud Electronica y Movil, David, pp. 1-8, 2024. https://doi.org/10.1109/AmITIC62658.2024.1074 7620

[16] Xi W., Wang E., and Shen Z., “Bring Old Mural Images to Life,” in Proceedings of the IEEE 7th Information Technology and Mechatronics Engineering Conference, Chongqing, pp. 933-937, 2023. https://doi.org/10.1109/ITOEC57671.2023.10291 602

[17] Xiao H., Zheng H., and Meng Q., “Research on Deep Learning-Driven High-Resolution Image Restoration for Murals from the Perspective of Vision Sensing,” IEEE Access, vol. 11, pp. 71472- 71483, 2023. https://doi.org/10.1109/ACCESS.2023.3295253

[18] Xu W. and Fu Y., “Deep Learning Algorithm in Ancient Relics Image Colour Restoration Technology,” Multimedia Tools Applications, vol. 82, pp. 23119-23150, 2023. https://doi.org/10.1007/s11042-022-14108-z

[19] Xu Z., Zhang X., Chen W., Liu J., and et al., “MuralDiff: Diffusion for Ancient Murals Restoration on Large-Scale Pre-Training,” IEEE Transactions on Emerging Topics in Deep Learning-Based Degradation Detection and Inpainting of Tamil Nadu Temple ... 353 Computational Intelligence, vol. 8, no. 3, pp. 2169-2181, 2024. https://doi.org/10.1109/TETCI.2024.3359038

[20] Yadav A., Sharma S., Sangwan D., and Kumar R., “Revitalizing Ancient Murals in the Shekhawati Region Through Image Inpainting Techniques,” in Proceedings of the 11th International Conference on Signal Processing and Integrated Networks, Noida, pp. 361-366, 2024. https://doi.org/10.1109/SPIN60856.2024.10511302

[21] Yu Z., Lyu S., Hou M., Sun Y., and Li L., “A New Method for Extracting Refined Sketches of Ancient Murals,” Sensors, vol. 24, no. 7, pp. 1-16, 2024. https://doi.org/10.3390/s24072213