{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,11]],"date-time":"2026-04-11T01:22:27Z","timestamp":1775870547269,"version":"3.50.1"},"reference-count":50,"publisher":"Association for Computing Machinery (ACM)","issue":"4","license":[{"start":{"date-parts":[[2019,7,12]],"date-time":"2019-07-12T00:00:00Z","timestamp":1562889600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"funder":[{"DOI":"10.13039\/501100002341","name":"Academy of Finland","doi-asserted-by":"crossref","award":["277833"],"award-info":[{"award-number":["277833"]}],"id":[{"id":"10.13039\/501100002341","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":["ACM Trans. Graph."],"published-print":{"date-parts":[[2019,8,31]]},"abstract":"<jats:p>It has been shown that rendering in the gradient domain, i.e., estimating finite difference gradients of image intensity using correlated samples, and combining them with direct estimates of pixel intensities by solving a screened Poisson problem, often offers fundamental benefits over merely sampling pixel intensities. The reasons can be traced to the frequency content of the light transport integrand and its interplay with the gradient operator. However, while they often yield state of the art performance among algorithms that are based on Monte Carlo sampling alone, gradient-domain rendering algorithms have, until now, not generally been competitive with techniques that combine Monte Carlo sampling with post-hoc noise removal using sophisticated non-linear filtering.<\/jats:p>\n          <jats:p>Drawing on the power of modern convolutional neural networks, we propose a novel reconstruction method for gradient-domain rendering. Our technique replaces the screened Poisson solver of previous gradient-domain techniques with a novel dense variant of the U-Net autoencoder, additionally taking auxiliary feature buffers as inputs. We optimize our network to minimize a perceptual image distance metric calibrated to the human visual system. Our results significantly improve the quality obtained from gradient-domain path tracing, allowing it to overtake state-of-the-art comparison techniques that denoise traditional Monte Carlo samplings. In particular, we observe that the correlated gradient samples --- that offer information about the smoothness of the integrand unavailable in standard Monte Carlo sampling --- notably improve image quality compared to an equally powerful neural model that does not make use of gradient samples.<\/jats:p>","DOI":"10.1145\/3306346.3323038","type":"journal-article","created":{"date-parts":[[2019,7,12]],"date-time":"2019-07-12T19:04:08Z","timestamp":1562958248000},"page":"1-12","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":37,"title":["Deep convolutional reconstruction for gradient-domain rendering"],"prefix":"10.1145","volume":"38","author":[{"given":"Markus","family":"Kettunen","sequence":"first","affiliation":[{"name":"Aalto University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Erik","family":"H\u00e4rk\u00f6nen","sequence":"additional","affiliation":[{"name":"Aalto University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jaakko","family":"Lehtinen","sequence":"additional","affiliation":[{"name":"Aalto University and Nvidia"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"320","published-online":{"date-parts":[[2019,7,12]]},"reference":[{"key":"e_1_2_1_1_1","volume-title":"Feature Generation for Adaptive Gradient-Domain Path Tracing. Computer Graphics Forum","author":"Back Jonghee","year":"2018"},{"key":"e_1_2_1_2_1","doi-asserted-by":"publisher","DOI":"10.1145\/3072959.3073708"},{"key":"e_1_2_1_3_1","doi-asserted-by":"publisher","DOI":"10.1145\/3130800.3130886"},{"key":"e_1_2_1_4_1","volume-title":"Proceedings of the 13th Eurographics Workshop on Rendering (EGRW '02)","author":"Bekaert Philippe","year":"2002"},{"key":"e_1_2_1_5_1","doi-asserted-by":"publisher","DOI":"10.1145\/3072959.3073698"},{"key":"e_1_2_1_6_1","doi-asserted-by":"publisher","DOI":"10.1111\/cgf.12954"},{"key":"e_1_2_1_7_1","doi-asserted-by":"publisher","DOI":"10.1137\/040616024"},{"key":"e_1_2_1_8_1","doi-asserted-by":"publisher","DOI":"10.1145\/3072959.3073601"},{"key":"e_1_2_1_9_1","doi-asserted-by":"publisher","DOI":"10.1145\/2366145.2366211"},{"key":"e_1_2_1_10_1","doi-asserted-by":"publisher","DOI":"10.1145\/3197517.3201363"},{"key":"e_1_2_1_11_1","volume-title":"Chitnis","author":"Guan Steven","year":"2018"},{"key":"e_1_2_1_12_1","doi-asserted-by":"publisher","DOI":"10.1145\/1661412.1618487"},{"key":"e_1_2_1_13_1","volume-title":"Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification. CoRR abs\/1502.01852","author":"He Kaiming","year":"2015"},{"key":"e_1_2_1_14_1","first-page":"2","article-title":"Gradient-Domain Photon Density","volume":"36","author":"Hua Binh-Son","year":"2017","journal-title":"Estimation. Comput. Graph. Forum"},{"key":"e_1_2_1_15_1","volume-title":"Weinberger","author":"Huang Gao","year":"2016"},{"key":"e_1_2_1_16_1","volume-title":"SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and &lt;1MB model size. CoRR abs\/1602.07360","author":"Iandola Forrest N.","year":"2016"},{"key":"e_1_2_1_17_1","unstructured":"Wenzel Jakob. 2010. Mitsuba renderer. http:\/\/www.mitsuba-renderer.org.  Wenzel Jakob. 2010. Mitsuba renderer. http:\/\/www.mitsuba-renderer.org."},{"key":"e_1_2_1_18_1","doi-asserted-by":"publisher","DOI":"10.1145\/2024156.2024215"},{"key":"e_1_2_1_19_1","volume-title":"The Beam Radiance Estimate for Volumetric Photon Mapping. Comput. Graph. Forum 27 (04","author":"Jarosz Wojciech","year":"2008"},{"key":"e_1_2_1_20_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46475-6_43"},{"key":"e_1_2_1_21_1","doi-asserted-by":"publisher","DOI":"10.1145\/15886.15902"},{"key":"e_1_2_1_22_1","doi-asserted-by":"publisher","DOI":"10.1145\/2766977"},{"key":"e_1_2_1_23_1","unstructured":"Markus Kettunen Erik H\u00e4rk\u00f6nen and Jaakko Lehtinen. 2019. Robust Perceptual Image Similarity via Self-Ensembled CNNs. Manuscript in preparation.  Markus Kettunen Erik H\u00e4rk\u00f6nen and Jaakko Lehtinen. 2019. Robust Perceptual Image Similarity via Self-Ensembled CNNs. Manuscript in preparation."},{"key":"e_1_2_1_24_1","doi-asserted-by":"publisher","DOI":"10.1145\/2766997"},{"key":"e_1_2_1_25_1","volume-title":"Kingma and Jimmy Ba","author":"Diederik","year":"2014"},{"key":"e_1_2_1_26_1","doi-asserted-by":"publisher","DOI":"10.1145\/2461912.2461943"},{"key":"e_1_2_1_27_1","volume-title":"Proceedings of the 35th International Conference on Machine Learning, PMLR","volume":"80","author":"Lehtinen Jaakko","year":"2018"},{"key":"e_1_2_1_28_1","volume-title":"H-DenseUNet: Hybrid Densely Connected UNet for Liver and Liver Tumor Segmentation from CT Volumes. CoRR abs\/1709.07330","author":"Li Xiaomeng","year":"2017"},{"key":"e_1_2_1_29_1","unstructured":"Andrew L. Maas. 2013. Rectifier Nonlinearities Improve Neural Network Acoustic Models.  Andrew L. Maas. 2013. Rectifier Nonlinearities Improve Neural Network Acoustic Models."},{"key":"e_1_2_1_30_1","volume-title":"Proc. Eurographics Symposium on Rendering.","author":"Manzi Marco","year":"2015"},{"key":"e_1_2_1_31_1","doi-asserted-by":"publisher","DOI":"10.1145\/2980179.2980256"},{"key":"e_1_2_1_32_1","volume-title":"Computer graphics forum","author":"Manzi Marco"},{"key":"e_1_2_1_33_1","volume-title":"Proc. NIPS.","author":"Mao Xiao-Jiao","year":"2016"},{"key":"e_1_2_1_34_1","doi-asserted-by":"publisher","DOI":"10.1145\/318009.318015"},{"key":"e_1_2_1_35_1","doi-asserted-by":"publisher","DOI":"10.1145\/2897824.2925936"},{"key":"e_1_2_1_36_1","volume-title":"An Introduction to Convolutional Neural Networks. CoRR abs\/1511.08458","author":"O'Shea Keiron","year":"2015"},{"key":"e_1_2_1_37_1","doi-asserted-by":"publisher","DOI":"10.1145\/566654.566575"},{"key":"e_1_2_1_38_1","volume-title":"U-Net: Convolutional Networks for Biomedical Image Segmentation. CoRR abs\/1505.04597","author":"Ronneberger Olaf","year":"2015"},{"key":"e_1_2_1_39_1","doi-asserted-by":"publisher","DOI":"10.1145\/2980179.2982443"},{"key":"e_1_2_1_40_1","doi-asserted-by":"publisher","DOI":"10.1145\/2070781.2024193"},{"key":"e_1_2_1_41_1","doi-asserted-by":"publisher","DOI":"10.1145\/2366145.2366214"},{"key":"e_1_2_1_42_1","doi-asserted-by":"publisher","DOI":"10.1111\/cgf.12219"},{"key":"e_1_2_1_43_1","volume-title":"Kingma","author":"Salimans Tim","year":"2016"},{"key":"e_1_2_1_44_1","volume-title":"Very Deep Convolutional Networks for Large-Scale Image Recognition. CoRR abs\/1409.1556","author":"Simonyan Karen","year":"2014"},{"key":"e_1_2_1_45_1","doi-asserted-by":"publisher","DOI":"10.2312\/sre.20171197"},{"key":"e_1_2_1_46_1","doi-asserted-by":"publisher","DOI":"10.1145\/218380.218498"},{"key":"e_1_2_1_47_1","doi-asserted-by":"publisher","DOI":"10.1145\/3197517.3201388"},{"key":"e_1_2_1_48_1","volume-title":"Calcium Removal From Cardiac CT Images Using Deep Convolutional Neural Network. CoRR abs\/1803.00399","author":"Yan Siming","year":"2018"},{"key":"e_1_2_1_49_1","volume-title":"The Unreasonable Effectiveness of Deep Features as a Perceptual Metric. CoRR abs\/1801.03924","author":"Zhang Richard","year":"2018"},{"key":"e_1_2_1_50_1","volume-title":"Yoon","author":"Zwicker Matthias","year":"2015"}],"container-title":["ACM Transactions on Graphics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3306346.3323038","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3306346.3323038","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,18]],"date-time":"2025-06-18T00:25:52Z","timestamp":1750206352000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3306346.3323038"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2019,7,12]]},"references-count":50,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2019,8,31]]}},"alternative-id":["10.1145\/3306346.3323038"],"URL":"https:\/\/doi.org\/10.1145\/3306346.3323038","relation":{},"ISSN":["0730-0301","1557-7368"],"issn-type":[{"value":"0730-0301","type":"print"},{"value":"1557-7368","type":"electronic"}],"subject":[],"published":{"date-parts":[[2019,7,12]]},"assertion":[{"value":"2019-07-12","order":2,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}