{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/towards-distortion-predictable-embedding-of","title":"Towards Distortion-Predictable Embedding of Neural Networks","arxiv_id":"1508.00102","date":"2015-08-01","proceeding":null,"authors":["Axel Angel"],"abstract":"Current research in Computer Vision has shown that Convolutional Neural\nNetworks (CNN) give state-of-the-art performance in many classification tasks\nand Computer Vision problems. The embedding of CNN, which is the internal\nrepresentation produced by the last layer, can indirectly learn topological and\nrelational properties. Moreover, by using a suitable loss function, CNN models\ncan learn invariance to a wide range of non-linear distortions such as\nrotation, viewpoint angle or lighting condition. In this work, new insights are\ndiscovered about CNN embeddings and a new loss function is proposed, derived\nfrom the contrastive loss, that creates models with more predicable mappings\nand also quantifies distortions. In typical distortion-dependent methods, there\nis no simple relation between the features corresponding to one image and the\nfeatures of this image distorted. Therefore, these methods require to\nfeed-forward inputs under every distortions in order to find the corresponding\nfeatures representations. Our contribution makes a step towards embeddings\nwhere features of distorted inputs are related and can be derived from each\nothers by the intensity of the distortion.","url_abs":"http://arxiv.org/abs/1508.00102v1","url_pdf":"http://arxiv.org/pdf/1508.00102v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"towards-distortion-predictable-embedding-of","repo_url":"https://github.com/axel-angel/master-project","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"caffe2","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}