{"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-high-resolution-video-generation-with","title":"Towards High Resolution Video Generation with Progressive Growing of Sliced Wasserstein GANs","arxiv_id":"1810.02419","date":"2018-10-04","proceeding":null,"authors":["Dinesh Acharya","Zhiwu Huang","Danda Pani Paudel","Luc van Gool"],"abstract":"The extension of image generation to video generation turns out to be a very\ndifficult task, since the temporal dimension of videos introduces an extra\nchallenge during the generation process. Besides, due to the limitation of\nmemory and training stability, the generation becomes increasingly challenging\nwith the increase of the resolution/duration of videos. In this work, we\nexploit the idea of progressive growing of Generative Adversarial Networks\n(GANs) for higher resolution video generation. In particular, we begin to\nproduce video samples of low-resolution and short-duration, and then\nprogressively increase both resolution and duration alone (or jointly) by\nadding new spatiotemporal convolutional layers to the current networks.\nStarting from the learning on a very raw-level spatial appearance and temporal\nmovement of the video distribution, the proposed progressive method learns\nspatiotemporal information incrementally to generate higher resolution videos.\nFurthermore, we introduce a sliced version of Wasserstein GAN (SWGAN) loss to\nimprove the distribution learning on the video data of high-dimension and\nmixed-spatiotemporal distribution. SWGAN loss replaces the distance between\njoint distributions by that of one-dimensional marginal distributions, making\nthe loss easier to compute. We evaluate the proposed model on our collected\nface video dataset of 10,900 videos to generate photorealistic face videos of\n256x256x32 resolution. In addition, our model also reaches a record inception\nscore of 14.57 in unsupervised action recognition dataset UCF-101.","url_abs":"http://arxiv.org/abs/1810.02419v2","url_pdf":"http://arxiv.org/pdf/1810.02419v2.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-high-resolution-video-generation-with","repo_url":"https://github.com/musikisomorphie/swd","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"action-recognition-in-videos","task_name":"Action Recognition"},{"task_slug":"image-generation","task_name":"Image Generation"},{"task_slug":"action-recognition","task_name":"Temporal Action Localization"},{"task_slug":"video-generation","task_name":"Video Generation"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1810.02419","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}