{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,25]],"date-time":"2026-06-25T17:43:34Z","timestamp":1782409414610,"version":"3.54.5"},"reference-count":17,"publisher":"Springer Science and Business Media LLC","issue":"4","license":[{"start":{"date-parts":[[2019,10,8]],"date-time":"2019-10-08T00:00:00Z","timestamp":1570492800000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2019,10,8]],"date-time":"2019-10-08T00:00:00Z","timestamp":1570492800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"funder":[{"DOI":"10.13039\/100010665","name":"H2020 Marie Sklodowska-Curie Actions","doi-asserted-by":"publisher","award":["642841"],"award-info":[{"award-number":["642841"]}],"id":[{"id":"10.13039\/100010665","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/100011199","name":"FP7 Ideas: European Research Council","doi-asserted-by":"publisher","award":["StG-2013-335373"],"award-info":[{"award-number":["StG-2013-335373"]}],"id":[{"id":"10.13039\/100011199","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000266","name":"Engineering and Physical Sciences Research Council","doi-asserted-by":"publisher","award":["EP\/K023578\/1"],"award-info":[{"award-number":["EP\/K023578\/1"]}],"id":[{"id":"10.13039\/501100000266","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Int J Comput Vis"],"published-print":{"date-parts":[[2020,4]]},"abstract":"<jats:title>Abstract<\/jats:title>\n<jats:p>Convolutional neural networks (CNNs) handle the case where filters extend beyond the image boundary using several heuristics, such as ,  or  padding. These schemes are applied in an ad-hoc fashion and, being weakly related to the image content and oblivious of the target task, result in low output quality at the boundary. In this paper, we propose a simple and effective improvement that <jats:italic>learns<\/jats:italic> the boundary handling itself. At training-time, the network is provided with a separate set of  boundary filters. At testing-time, we use these filters which have learned to extrapolate features at the boundary in an optimal way for the specific task. Our extensive evaluation, over a wide range of architectural changes (variations of layers, feature channels, or both), shows how the  filters result in improved boundary handling. Furthermore, we investigate the efficacy of variations of such boundary filters with respect to convergence speed and accuracy. Finally, we demonstrate an improvement of 5\u201320% across the board of typical CNN applications (colorization, de-Bayering, optical flow, disparity estimation, and super-resolution). Supplementary material and code can be downloaded from the project page:\u00a0<jats:ext-link xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" ext-link-type=\"uri\" xlink:href=\"http:\/\/geometry.cs.ucl.ac.uk\/projects\/2019\/investigating-edge\/\">http:\/\/geometry.cs.ucl.ac.uk\/projects\/2019\/investigating-edge\/<\/jats:ext-link>.<\/jats:p>","DOI":"10.1007\/s11263-019-01223-y","type":"journal-article","created":{"date-parts":[[2019,10,8]],"date-time":"2019-10-08T08:03:11Z","timestamp":1570521791000},"page":"773-782","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":33,"title":["Learning on the Edge: Investigating Boundary Filters in CNNs"],"prefix":"10.1007","volume":"128","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-8431-4335","authenticated-orcid":false,"given":"Carlo","family":"Innamorati","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tobias","family":"Ritschel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tim","family":"Weyrich","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Niloy J.","family":"Mitra","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2019,10,8]]},"reference":[{"key":"1223_CR1","unstructured":"Chollet, F., et al. (2015). Keras. Retrieved July, 2019, from \nhttps:\/\/keras.io"},{"key":"1223_CR2","unstructured":"Ciregan, D., Meier, U., & Schmidhuber, J. (2012). Multi-column deep neural networks for image classification. In CVPR (pp. 3642\u201349)."},{"key":"1223_CR3","unstructured":"Cohen, T., & Welling, M. (2016). Group equivariant convolutional networks. In ICML (pp. 2990\u20132999)."},{"key":"1223_CR4","doi-asserted-by":"crossref","unstructured":"Deng, J., Dong, W., Socher, R., Jia Li, L., Li, K., & Fei-fei, L. (2009). Imagenet: A large-scale hierarchical image database. In CVPR.","DOI":"10.1109\/CVPR.2009.5206848"},{"issue":"2","key":"1223_CR5","doi-asserted-by":"publisher","first-page":"295","DOI":"10.1109\/TPAMI.2015.2439281","volume":"38","author":"C Dong","year":"2016","unstructured":"Dong, C., Loy, C. C., He, K., & Tang, X. (2016). Image super-resolution using deep convolutional networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 38(2), 295\u2013307. \nhttps:\/\/doi.org\/10.1109\/TPAMI.2015.2439281\n\n.","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"1223_CR6","doi-asserted-by":"crossref","unstructured":"Dosovitskiy, A., Fischer, P., Ilg, E., H\u00e4usser, P., Haz\u0131rba\u015f, C., Golkov, V., v.d. Smagt, P., Cremers, D., & Brox, T. (2015). Flownet: Learning optical flow with convolutional networks. In ICCV.","DOI":"10.1109\/ICCV.2015.316"},{"key":"1223_CR7","unstructured":"Fukushima, K., & Miyake, S. (1982). Neocognitron: A self-organizing neural network model for a mechanism of visual pattern recognition. In Competition and cooperation in neural nets (pp. 267\u201385). Berlin: Springer."},{"issue":"6","key":"1223_CR8","doi-asserted-by":"publisher","first-page":"191","DOI":"10.1145\/2980179.2982399","volume":"35","author":"M Gharbi","year":"2016","unstructured":"Gharbi, M., Chaurasia, G., Paris, S., & Durand, F. (2016). Deep joint demosaicking and denoising. ACM Transactions on Graphics, 35(6), 191.","journal-title":"ACM Transactions on Graphics"},{"key":"1223_CR9","volume-title":"Deep learning","author":"I Goodfellow","year":"2016","unstructured":"Goodfellow, I., Bengio, Y., Courville, A., & Bengio, Y. (2016). Deep learning (Vol. 1). Cambridge: MIT Press."},{"key":"1223_CR10","unstructured":"Innamorati, C., Ritschel, T., Weyrich, T., & Mitra, N. J. (2018). Learning on the edge: Explicit boundary handling in CNNs. In Proceedings of the british machine vision conference (BMVC). BMVA Press. Selected for oral presentation"},{"key":"1223_CR11","unstructured":"Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. In NIPS (pp. 1097\u20131105)."},{"key":"1223_CR12","unstructured":"Ren, J.S., Xu, L., Yan, Q., & Sun, W. (2015). Shepard convolutional neural networks. In NIPS (pp. 901\u201309)."},{"key":"1223_CR13","doi-asserted-by":"crossref","unstructured":"Ronneberger, O., Fischer, P., & Brox, T. (2015). U-net: Convolutional networks for biomedical image segmentation. In International conference on medical image computing and computer-assisted intervention (pp. 234\u201341).","DOI":"10.1007\/978-3-319-24574-4_28"},{"issue":"3","key":"1223_CR14","doi-asserted-by":"publisher","first-page":"211","DOI":"10.1007\/s11263-015-0816-y","volume":"115","author":"O Russakovsky","year":"2015","unstructured":"Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., et al. (2015). ImageNet large scale visual recognition challenge. International Journal of Computer Vision (IJCV), 115(3), 211\u2013252. \nhttps:\/\/doi.org\/10.1007\/s11263-015-0816-y\n\n.","journal-title":"International Journal of Computer Vision (IJCV)"},{"key":"1223_CR15","unstructured":"Uhrig, J., Schneider, N., Schneider, L., Franke, U., Brox, T., & Geiger, A. (2017). Sparsity invariant CNNs. \narXiv:1708.06500\n\n."},{"key":"1223_CR16","doi-asserted-by":"crossref","unstructured":"Worrall, D. E., Garbin, S. J., Turmukhambetov, D., & Brostow, G. J. (2017). Harmonic networks: Deep translation and rotation equivariance. In CVPR.","DOI":"10.1109\/CVPR.2017.758"},{"key":"1223_CR17","doi-asserted-by":"crossref","unstructured":"Zhang, R., Isola, P., & Efros, A. A. (2016). Colorful image colorization. In ECCV.","DOI":"10.1007\/978-3-319-46487-9_40"}],"container-title":["International Journal of Computer Vision"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/link.springer.com\/content\/pdf\/10.1007\/s11263-019-01223-y.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"http:\/\/link.springer.com\/article\/10.1007\/s11263-019-01223-y\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"http:\/\/link.springer.com\/content\/pdf\/10.1007\/s11263-019-01223-y.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2020,10,6]],"date-time":"2020-10-06T23:12:02Z","timestamp":1602025922000},"score":1,"resource":{"primary":{"URL":"http:\/\/link.springer.com\/10.1007\/s11263-019-01223-y"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2019,10,8]]},"references-count":17,"journal-issue":{"issue":"4","published-print":{"date-parts":[[2020,4]]}},"alternative-id":["1223"],"URL":"https:\/\/doi.org\/10.1007\/s11263-019-01223-y","relation":{},"ISSN":["0920-5691","1573-1405"],"issn-type":[{"value":"0920-5691","type":"print"},{"value":"1573-1405","type":"electronic"}],"subject":[],"published":{"date-parts":[[2019,10,8]]},"assertion":[{"value":"19 January 2019","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"27 August 2019","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"8 October 2019","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}}]}}