{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,7]],"date-time":"2026-05-07T02:40:55Z","timestamp":1778121655153,"version":"3.51.4"},"reference-count":48,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2024,11,6]],"date-time":"2024-11-06T00:00:00Z","timestamp":1730851200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Artif. Intell."],"abstract":"<jats:p>Mural paintings are vital cultural expressions, enriching our lives by beautifying spaces, conveying messages, telling stories, and evoking emotions. Ancient temple murals degrade over time due to natural aging, physical damage, etc. Preserving these cultural treasures is challenging. Image inpainting is often used for digital restoration, but existing methods typically overlook naturally degraded areas, using randomly generated binary masks or small, narrow regions for repair. This study proposes a novel architecture to reconstruct large areas of naturally degraded murals, maintaining intrinsic details, avoiding color bias, and preserving artistic excellence. The architecture integrates generative adversarial networks (GANs) and the diffusion model, including a whole structure formation network (WSFN), a semantic color network (SCN), and a diffusion mixture distribution (DIMD) discriminator. The WSFN uses the original image, a line drawing, and an edge map to capture mural details, which are then texturally inpainted in the SCN using gated convolution for enhanced results. Special attention is given to globally extending the receptive field for large-area inpainting. The model is evaluated using custom-degraded mural images collected from Tamil Nadu temples. Quantitative analysis showed superior results than state-of-the-art methods, with SSIM, MSE, PSNR, and LPIPS values of 0.8853, 0.0021, 29.8826, and 0.0426, respectively.<\/jats:p>","DOI":"10.3389\/frai.2024.1453847","type":"journal-article","created":{"date-parts":[[2024,11,6]],"date-time":"2024-11-06T06:11:56Z","timestamp":1730873516000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":8,"title":["Inpainting of damaged temple murals using edge- and line-guided diffusion patch GAN"],"prefix":"10.3389","volume":"7","author":[{"given":"G.","family":"Sumathi","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"M.","family":"Uma Devi","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2024,11,6]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"469","DOI":"10.1080\/00401706.1971.10488811","article-title":"Mean square error of prediction as a criterion for selecting variables","volume":"13","author":"Allen","year":"1971","journal-title":"Technometrics"},{"key":"ref2","doi-asserted-by":"publisher","first-page":"2481","DOI":"10.1109\/TPAMI.2016.2644615","article-title":"Segnet: a deep convolutional encoder-decoder architecture for image segmentation","volume":"39","author":"Badrinarayanan","year":"2017","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref3","doi-asserted-by":"publisher","first-page":"24","DOI":"10.1145\/1531326.1531330","article-title":"PatchMatch: a randomized correspondence algorithm for structural image editing","volume":"28","author":"Barnes","year":"2009","journal-title":"ACM Trans. Graph."},{"key":"ref4","doi-asserted-by":"publisher","first-page":"103883","DOI":"10.1016\/j.cviu.2023.103883","article-title":"MFMAM: image inpainting via multi-scale feature module with attention module","volume":"238","author":"Chen","year":"2024","journal-title":"Comput. Vis. Image Underst."},{"key":"ref5","doi-asserted-by":"publisher","first-page":"103776","DOI":"10.1016\/j.jvcir.2023.103776","article-title":"FFTI: image inpainting algorithm via features fusion and two-steps inpainting","volume":"91","author":"Chen","year":"2023","journal-title":"J. Vis. Commun. Image Represent."},{"key":"ref6","doi-asserted-by":"publisher","first-page":"105814","DOI":"10.1016\/j.engappai.2022.105814","article-title":"RFA-net: residual feature attention network for fine-grained image inpainting","volume":"119","author":"Chen","year":"2023","journal-title":"Eng. Appl. Artif. Intell."},{"key":"ref7","doi-asserted-by":"publisher","first-page":"032040","DOI":"10.1088\/1742-6596\/1302\/3\/032040","article-title":"Image inpainting for digital Dunhuang murals using partial convolutions and sliding window method","volume":"1302","author":"Chen","year":"2019","journal-title":"J. Phys. Conf. Ser."},{"key":"ref8","doi-asserted-by":"publisher","first-page":"2091","DOI":"10.3390\/s21062091","article-title":"Color-balanced edge-guided digital inpainting: applications on artworks","volume":"21","author":"Ciortan","year":"2021","journal-title":"Sensors"},{"key":"ref9","volume-title":"Patch-based image inpainting with generative adversarial networks","author":"Demir","year":"2018"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"103","DOI":"10.1145\/2830541","article-title":"What makes Paris look like Paris?","volume":"58","author":"Doersch","year":"2015","journal-title":"Commun. ACM"},{"key":"ref11","author":"Gatys","year":"2016"},{"key":"ref12","author":"Guo","year":"2021"},{"key":"ref13","author":"He","year":"2016"},{"key":"ref14","author":"Hore","year":"2010"},{"key":"ref15","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3072959.3073659","article-title":"Globally and locally consistent image completion","volume":"36","author":"Iizuka","year":"2017","journal-title":"ACM Trans. Graphics"},{"key":"ref16","author":"Isola","year":"2017"},{"key":"ref17","author":"Jaidilert","year":"2018"},{"key":"ref18","doi-asserted-by":"publisher","first-page":"84360","DOI":"10.1109\/ACCESS.2023.3299876","article-title":"Image inpainting of thangka murals using edge-assisted feature fusion and self attention based local refine network","volume":"11","author":"Jia","year":"2023","journal-title":"IEEE Access"},{"key":"ref19","author":"Li","year":"2020"},{"key":"ref20","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1007\/s11427-017-9265-8","article-title":"Dating ancient paintings of Mogao grottoes using deeply learnt visual codes, science China","volume":"61","author":"Li","year":"2018","journal-title":"Inf. Sci."},{"key":"ref21","volume-title":"Line drawing guided progressive Inpainting of mural damages","author":"Li","year":"2022"},{"key":"ref22","doi-asserted-by":"publisher","first-page":"4303","DOI":"10.1007\/s00371-023-03083-7","article-title":"Guiding image inpainting via structure and texture features with dual encoder","volume":"40","author":"Lian","year":"2024","journal-title":"Vis. Comput."},{"key":"ref23","doi-asserted-by":"publisher","first-page":"5830","DOI":"10.1109\/CVPR.2019.00598","article-title":"Sketchgan: joint sketch completion and recognition with generative adversarial network","author":"Liu","year":"2019","journal-title":"CVPR"},{"key":"ref24","doi-asserted-by":"publisher","first-page":"3730","DOI":"10.1109\/ICCV.2015.425","article-title":"Deep learning face attributes in the wild","volume":"12","author":"Liu","year":"2015","journal-title":"ICCV"},{"key":"ref25","author":"Liu","year":"2018"},{"key":"ref26","volume-title":"Edgeconnect: Generative image inpainting with adversarial edge learning, ICCV","author":"Nazeri","year":"2019"},{"key":"ref27","author":"Pathak","year":"2016"},{"key":"ref28","author":"Poma","year":"2020"},{"key":"ref29","doi-asserted-by":"publisher","first-page":"39","DOI":"10.1186\/s40494-024-01159-8","article-title":"Dunhuang murals image restoration method based on generative adversarial network","volume":"12","author":"Ren","year":"2024","journal-title":"Herit. Sci."},{"key":"ref30","volume-title":"Stable and controllable neural texture synthesis and style transfer using histogram losses","author":"Risser","year":"2017"},{"key":"ref31","author":"Song","year":"2020"},{"key":"ref32","doi-asserted-by":"publisher","first-page":"9780","DOI":"10.3390\/s22249780","article-title":"Enhancement and restoration of scratched murals based on hyperspectral imaging\u2014a case study of murals in the Baoguang hall of Qutan Temple, Qinghai, China","volume":"22","author":"Sun","year":"2022","journal-title":"Sensors"},{"key":"ref33","doi-asserted-by":"publisher","first-page":"266","DOI":"10.1186\/s40494-024-01391-2","article-title":"Ancient paintings inpainting based on dual encoders and contextual information","volume":"12","author":"Sun","year":"2024","journal-title":"Herit. Sci."},{"key":"ref34","doi-asserted-by":"publisher","first-page":"600","DOI":"10.1109\/TIP.2003.819861","article-title":"Image quality assessment: from error visibility to structural similarity","volume":"13","author":"Wang","year":"2004","journal-title":"IEEE Trans. Image Process."},{"key":"ref35","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3280790","article-title":"Inpainting of Dunhuang murals by sparsely modeling the texture similarity and structure continuity","volume":"12","author":"Wang","year":"2019","journal-title":"J. Comput. Cult. Heritage"},{"key":"ref36","author":"Wang","year":"2018"},{"key":"ref37","author":"Wang","year":"2020"},{"key":"ref38","doi-asserted-by":"publisher","first-page":"3720","DOI":"10.1109\/TIP.2021.3064268","article-title":"Thanka mural inpainting based on multi-scale adaptive partial convolution and stroke like mask","volume":"30","author":"Wang","year":"2021","journal-title":"IEEE Trans. Image Process."},{"key":"ref39","doi-asserted-by":"publisher","first-page":"3972","DOI":"10.3390\/app13063972","article-title":"Fragments in painting for tomb murals using a dual-attention mechanism GAN with improved generators","volume":"13","author":"Wu","year":"2023","journal-title":"Appl. Sci."},{"key":"ref40","doi-asserted-by":"publisher","first-page":"169","DOI":"10.1186\/s40494-023-01015-1","article-title":"Digital inpainting of mural images based on DC-CycleGAN","volume":"11","author":"Xu","year":"2023","journal-title":"Herit. Sci."},{"key":"ref41","author":"Yan","year":"2018"},{"key":"ref42","author":"Yang","year":"2017"},{"key":"ref43","doi-asserted-by":"publisher","first-page":"7508","DOI":"10.1109\/CVPR42600.2020.00753","article-title":"Contextual residual aggregation for ultra high-resolution image in painting","author":"Yi","year":"2020","journal-title":"CVPR"},{"key":"ref44","author":"Yu","year":"2018"},{"key":"ref45","author":"Yu","year":"2019"},{"key":"ref46","author":"Zhang","year":"2018"},{"key":"ref47","doi-asserted-by":"publisher","first-page":"1452","DOI":"10.1109\/TPAMI.2017.2723009","article-title":"Places: a 10 million image database for scene recognition","volume":"40","author":"Zhou","year":"2017","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref48","doi-asserted-by":"publisher","first-page":"146","DOI":"10.1016\/j.patrec.2014.07.002","article-title":"Chronological classification of ancient paintings using appearance and shape features","volume":"49","author":"Zou","year":"2014","journal-title":"Pattern Recogn. Lett."}],"container-title":["Frontiers in Artificial Intelligence"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2024.1453847\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,11,6]],"date-time":"2024-11-06T06:11:59Z","timestamp":1730873519000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/frai.2024.1453847\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,11,6]]},"references-count":48,"alternative-id":["10.3389\/frai.2024.1453847"],"URL":"https:\/\/doi.org\/10.3389\/frai.2024.1453847","relation":{},"ISSN":["2624-8212"],"issn-type":[{"value":"2624-8212","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,11,6]]},"article-number":"1453847"}}