{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,27]],"date-time":"2026-05-27T23:27:32Z","timestamp":1779924452700,"version":"3.53.1"},"reference-count":26,"publisher":"Springer Science and Business Media LLC","issue":"1","license":[{"start":{"date-parts":[[2021,7,2]],"date-time":"2021-07-02T00:00:00Z","timestamp":1625184000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2021,7,2]],"date-time":"2021-07-02T00:00:00Z","timestamp":1625184000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/501100002790","name":"Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100002790","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["J Big Data"],"published-print":{"date-parts":[[2021,12]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>Large data requirements are often the main hurdle in training neural networks. Convolutional neural network (CNN) classifiers in particular require tens of thousands of pre-labeled images per category to approach human-level accuracy, while often failing to generalized to out-of-domain test sets. The acquisition and labelling of such datasets is often an expensive, time consuming and tedious task in practice. Synthetic data provides a cheap and efficient solution to assemble such large datasets. Using domain randomization (DR), we show that a sufficiently well generated synthetic image dataset can be used to train a neural network classifier that rivals state-of-the-art models trained on real datasets, achieving accuracy levels as high as 88% on a baseline cats vs dogs classification task. We show that the most important domain randomization parameter is a large variety of subjects, while secondary parameters such as lighting and textures are found to be less significant to the model accuracy. Our results also provide evidence to suggest that models trained on domain randomized images transfer to new domains better than those trained on real photos. Model performance appears to remain stable as the number of categories increases.<\/jats:p>","DOI":"10.1186\/s40537-021-00455-5","type":"journal-article","created":{"date-parts":[[2021,7,2]],"date-time":"2021-07-02T07:03:29Z","timestamp":1625209409000},"update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":23,"title":["Domain randomization for neural network classification"],"prefix":"10.1186","volume":"8","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6149-138X","authenticated-orcid":false,"given":"Svetozar Zarko","family":"Valtchev","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jianhong","family":"Wu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2021,7,2]]},"reference":[{"key":"455_CR1","doi-asserted-by":"publisher","unstructured":"Wabartha M, Durand A, Fran\u0107ois-Lavet V, Pineau J. Handling black swan events in deep learning with diversely extrapolated neural networks. In: Bessiere C, editor. Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence, IJCAI-20. International Joint Conferences on Artificial Intelligence Organization; 2020. p. 2140\u20132147. Main track. Available from: https:\/\/doi.org\/10.24963\/ijcai.2020\/296.","DOI":"10.24963\/ijcai.2020\/296"},{"key":"455_CR2","doi-asserted-by":"crossref","unstructured":"Tremblay J, Prakash A, Acuna D, Brophy M, Jampani V, Anil C, et\u00a0al. Training deep networks with synthetic data: bridging the reality gap by domain randomization. In: 2018 IEEE\/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW); 2018. p. 1082\u201310828.","DOI":"10.1109\/CVPRW.2018.00143"},{"key":"455_CR3","doi-asserted-by":"crossref","unstructured":"Shafaei A, Little JJ, Schmidt M. Play and learn: using video games to train computer vision models; 2016;26.1\u201326.13.","DOI":"10.5244\/C.30.26"},{"key":"455_CR4","doi-asserted-by":"crossref","unstructured":"Atapour-Abarghouei A, Breckon TP. Real-time monocular depth estimation using synthetic data with domain adaptation via image style transfer. In: 2018 IEEE\/CVF Conference on Computer Vision and Pattern Recognition; 2018. p. 2800\u20132810.","DOI":"10.1109\/CVPR.2018.00296"},{"key":"455_CR5","doi-asserted-by":"crossref","unstructured":"Shermeyer J, Hossler T, Etten AV, Hogan D, Lewis R, Kim D. RarePlanes: synthetic data takes flight; 2020.","DOI":"10.1109\/WACV48630.2021.00025"},{"key":"455_CR6","doi-asserted-by":"crossref","unstructured":"Yue X, Zhang Y, Zhao S, Sangiovanni-Vincentelli A, Keutzer K, Gong B. Domain randomization and pyramid consistency: Simulation-to-Real Generalization Without Accessing Target Domain Data. In: 2019 IEEE\/CVF International Conference on Computer Vision (ICCV); 2019. p. 2100\u20132110.","DOI":"10.1109\/ICCV.2019.00219"},{"key":"455_CR7","doi-asserted-by":"crossref","unstructured":"Tobin J, Fong R, Ray A, Schneider J, Zaremba W, Abbeel P. Domain randomization for transferring deep neural networks from simulation to the real world. In: 2017 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS); 2017. p. 23\u201330.","DOI":"10.1109\/IROS.2017.8202133"},{"key":"455_CR8","doi-asserted-by":"publisher","first-page":"202","DOI":"10.1007\/978-3-319-49409-8_18","volume-title":"Computer vision-ECCV 2016 workshops","author":"Y Movshovitz-Attias","year":"2016","unstructured":"Movshovitz-Attias Y, Kanade T, Sheikh Y. How useful is photo-realistic rendering for visual learning? In: Hua G, J\u00e9gou H, editors. Computer vision-ECCV 2016 workshops. Cham: Springer International Publishing; 2016. p. 202\u201317."},{"key":"455_CR9","doi-asserted-by":"crossref","unstructured":"Tobin J, Biewald L, Duan R, Andrychowicz M, Handa A, Kumar V, et\u00a0al. Domain randomization and generative models for robotic grasping. In: 2018 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS); 2018. p. 3482\u20133489.","DOI":"10.1109\/IROS.2018.8593933"},{"issue":"1","key":"455_CR10","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/TRO.2019.2942989","volume":"36","author":"A Loquercio","year":"2020","unstructured":"Loquercio A, Kaufmann E, Ranftl R, Dosovitskiy A, Koltun V, Scaramuzza D. Deep drone racing: from simulation to reality with domain randomization. IEEE Trans Robot. 2020;36(1):1\u201314. https:\/\/doi.org\/10.1109\/TRO.2019.2942989.","journal-title":"IEEE Trans Robot."},{"key":"455_CR11","doi-asserted-by":"publisher","first-page":"682","DOI":"10.1007\/978-3-030-11009-3_42","volume-title":"Computer vision-ECCV 2018 workshops","author":"S Hinterstoisser","year":"2019","unstructured":"Hinterstoisser S, Lepetit V, Wohlhart P, Konolige K. On pre-trained image features and synthetic images for deep learning. In: Leal-Taix\u00e9 L, Roth S, editors. Computer vision-ECCV 2018 workshops. Cham: Springer International Publishing; 2019. p. 682\u201397."},{"key":"455_CR12","unstructured":"Peng X, Sun B, Ali K, Saenko K. Exploring invariances in deep convolutional neural networks using synthetic images. CoRR. 2014;abs\/1412.7122. Available from: arxiv.org\/abs\/1412.7122."},{"key":"455_CR13","doi-asserted-by":"crossref","unstructured":"Chen W, Wang H, Li Y, Su H, Wang Z, Tu C, et\u00a0al. Synthesizing training images for boosting human 3D Pose Estimation. In: 2016 Fourth International Conference on 3D Vision (3DV); 2016. p. 479\u2013488.","DOI":"10.1109\/3DV.2016.58"},{"key":"455_CR14","doi-asserted-by":"publisher","first-page":"517","DOI":"10.1007\/978-3-319-24947-6_43","volume-title":"Pattern Recog.","author":"B Pepik","year":"2015","unstructured":"Pepik B, Benenson R, Ritschel T, Schiele B. What is holding back convnets for detection? In: Gall J, Gehler P, Leibe B, editors. Pattern Recog. Cham: Springer International Publishing; 2015. p. 517\u201328."},{"key":"455_CR15","doi-asserted-by":"crossref","unstructured":"Peng X, Sun B, Ali K, Saenko K. Learning deep object detectors from 3D models. In: 2015 IEEE International Conference on Computer Vision (ICCV); 2015. p. 1278\u20131286.","DOI":"10.1109\/ICCV.2015.151"},{"key":"455_CR16","doi-asserted-by":"crossref","unstructured":"Mozian M, Camilo Gamboa Higuera J, Meger D, Dudek G. Learning domain randomization distributions for training robust locomotion policies. In: 2020 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS); 2020. p. 6112\u20136117.","DOI":"10.1109\/IROS45743.2020.9341019"},{"key":"455_CR17","unstructured":"Mehta B, Diaz M, Golemo F, Pal CJ, Paull L. Active domain randomization; 2019."},{"key":"455_CR18","doi-asserted-by":"crossref","unstructured":"Prakash A, Boochoon S, Brophy M, Acuna D, Cameracci E, State G, et\u00a0al. Structured domain randomization: bridging the reality gap by context-aware synthetic data. 2019. p. 7249\u20137255.","DOI":"10.1109\/ICRA.2019.8794443"},{"key":"455_CR19","unstructured":"Open AI, Akkaya. I, Andrychowicz M, Chociej M, Litwin M, McGrew B, et al. Solving rubik\u2019s cube with a robot hand; 2019."},{"key":"455_CR20","doi-asserted-by":"publisher","unstructured":"Schraml D. Physically based synthetic image generation for machine learning: a review of pertinent literature. In: Rosenberger M, Dittrich PG, Zagar B, editors. Photonics and education in measurement science 2019. vol. 11144. International Society for Optics and Photonics. SPIE; 2019. p. 108 \u2013 120. Available from: https:\/\/doi.org\/10.1117\/12.2533485.","DOI":"10.1117\/12.2533485"},{"key":"455_CR21","unstructured":"Kaggle. Dogs vs cats dataset; 2013. Available from: https:\/\/www.kaggle.com\/c\/dogs-vs-cats\/data."},{"key":"455_CR22","doi-asserted-by":"crossref","unstructured":"Selvaraju RR, Cogswell M, Das A, Vedantam R, Parikh D, Batra D. Grad-CAM: visual explanations from deep networks via gradient-based localization. In: 2017 IEEE International Conference on Computer Vision (ICCV); 2017. p. 618\u2013626.","DOI":"10.1109\/ICCV.2017.74"},{"key":"455_CR23","doi-asserted-by":"publisher","first-page":"818","DOI":"10.1007\/978-3-319-10590-1_53","volume-title":"Computer vision-ECCV 2014","author":"MD Zeiler","year":"2014","unstructured":"Zeiler MD, Fergus R. Visualizing and understanding convolutional networks. In: Fleet D, Pajdla T, Schiele B, Tuytelaars T, editors. Computer vision-ECCV 2014. Cham: Springer International Publishing; 2014. p. 818\u201333."},{"key":"455_CR24","unstructured":"Opelt A, Pinz A. GRAZ-02 Dataset; 2004. Available from: https:\/\/www-old.emt.tugraz.at\/~pinz\/data\/GRAZ_02\/."},{"key":"455_CR25","unstructured":"Khan S. Bicycle image dataset; 2020. Available from: http:\/\/maviintelligence.com\/bicycle-image-dataset\/."},{"key":"455_CR26","doi-asserted-by":"crossref","unstructured":"Krause J, Stark M, Deng J, Fei-Fei L. 3D object representations for fine-grained categorization. In: 4th International IEEE Workshop on 3D Representation and Recognition (3dRR-13). Sydney, Australia; 2013.","DOI":"10.1109\/ICCVW.2013.77"}],"container-title":["Journal of Big Data"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s40537-021-00455-5.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1186\/s40537-021-00455-5\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1186\/s40537-021-00455-5.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2021,7,2]],"date-time":"2021-07-02T07:08:32Z","timestamp":1625209712000},"score":1,"resource":{"primary":{"URL":"https:\/\/journalofbigdata.springeropen.com\/articles\/10.1186\/s40537-021-00455-5"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,7,2]]},"references-count":26,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2021,12]]}},"alternative-id":["455"],"URL":"https:\/\/doi.org\/10.1186\/s40537-021-00455-5","relation":{"has-preprint":[{"id-type":"doi","id":"10.21203\/rs.3.rs-75530\/v1","asserted-by":"object"}]},"ISSN":["2196-1115"],"issn-type":[{"value":"2196-1115","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,7,2]]},"assertion":[{"value":"14 September 2020","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"3 May 2021","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"2 July 2021","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"Not applicable","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethics approval and consent to participate"}},{"value":"Not applicable","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Consent for publication"}},{"value":"The synthetic datasets generated during the current study have been made available at the Synthetic Image Dataset (Cats, Dogs, Bikes, Cars) Kaggle repository at\n                      \n                      . The real image datasets used and\/or analysed are available in the Dogs vs. Cats Kaggle, GRAZ-02, Maviintelligence and StanfordAI repositories [\n                      \n                      ,\n                      \n                      \u2013\n                      \n                      ], at\n                      \n                      ,\n                      \n                      ,\n                      \n                      , and\n                      \n                      respectively.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Availability of data and materials"}},{"value":"The authors declare that they have no competing interests.","order":5,"name":"Ethics","group":{"name":"EthicsHeading","label":"Competing interests"}}],"article-number":"94"}}