Scaling Mesh Generation via Compressive Tokenization
We propose a compressive yet effective mesh representation, Blocked and Patchified Tokenization (BPT), facilitating the generation of meshes exceeding 8k faces. BPT compresses mesh sequences by employing block-wise indexing and patch aggregation, reducing their length by approximately 75\% compared...
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , |
|---|---|
| Format: | Artikel |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext bestellen |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | |
|---|---|
| container_issue | |
| container_start_page | |
| container_title | |
| container_volume | |
| creator | Weng, Haohan Zhao, Zibo Lei, Biwen Yang, Xianghui Liu, Jian Lai, Zeqiang Chen, Zhuo Liu, Yuhong Jiang, Jie Guo, Chunchao Zhang, Tong Gao, Shenghua Chen, C. L. Philip |
| description | We propose a compressive yet effective mesh representation, Blocked and
Patchified Tokenization (BPT), facilitating the generation of meshes exceeding
8k faces. BPT compresses mesh sequences by employing block-wise indexing and
patch aggregation, reducing their length by approximately 75\% compared to the
original sequences. This compression milestone unlocks the potential to utilize
mesh data with significantly more faces, thereby enhancing detail richness and
improving generation robustness. Empowered with the BPT, we have built a
foundation mesh generative model training on scaled mesh data to support
flexible control for point clouds and images. Our model demonstrates the
capability to generate meshes with intricate details and accurate topology,
achieving SoTA performance on mesh generation and reaching the level for direct
product usage. |
| doi_str_mv | 10.48550/arxiv.2411.07025 |
| format | Article |
| fullrecord | <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2411_07025</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2411_07025</sourcerecordid><originalsourceid>FETCH-arxiv_primary_2411_070253</originalsourceid><addsrcrecordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjE01DMwNzAy5WQwCU5OzMnMS1fwTS3OUHBPzUstSizJzM9TKMtMVHDOzy0oSi0uzixLVQjJz07Ny6wCS_IwsKYl5hSn8kJpbgZ5N9cQZw9dsPnxBUWZuYlFlfEge-LB9hgTVgEAOAoycw</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Scaling Mesh Generation via Compressive Tokenization</title><source>arXiv.org</source><creator>Weng, Haohan ; Zhao, Zibo ; Lei, Biwen ; Yang, Xianghui ; Liu, Jian ; Lai, Zeqiang ; Chen, Zhuo ; Liu, Yuhong ; Jiang, Jie ; Guo, Chunchao ; Zhang, Tong ; Gao, Shenghua ; Chen, C. L. Philip</creator><creatorcontrib>Weng, Haohan ; Zhao, Zibo ; Lei, Biwen ; Yang, Xianghui ; Liu, Jian ; Lai, Zeqiang ; Chen, Zhuo ; Liu, Yuhong ; Jiang, Jie ; Guo, Chunchao ; Zhang, Tong ; Gao, Shenghua ; Chen, C. L. Philip</creatorcontrib><description>We propose a compressive yet effective mesh representation, Blocked and
Patchified Tokenization (BPT), facilitating the generation of meshes exceeding
8k faces. BPT compresses mesh sequences by employing block-wise indexing and
patch aggregation, reducing their length by approximately 75\% compared to the
original sequences. This compression milestone unlocks the potential to utilize
mesh data with significantly more faces, thereby enhancing detail richness and
improving generation robustness. Empowered with the BPT, we have built a
foundation mesh generative model training on scaled mesh data to support
flexible control for point clouds and images. Our model demonstrates the
capability to generate meshes with intricate details and accurate topology,
achieving SoTA performance on mesh generation and reaching the level for direct
product usage.</description><identifier>DOI: 10.48550/arxiv.2411.07025</identifier><language>eng</language><subject>Computer Science - Computer Vision and Pattern Recognition ; Computer Science - Graphics</subject><creationdate>2024-11</creationdate><rights>http://creativecommons.org/licenses/by/4.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2411.07025$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2411.07025$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Weng, Haohan</creatorcontrib><creatorcontrib>Zhao, Zibo</creatorcontrib><creatorcontrib>Lei, Biwen</creatorcontrib><creatorcontrib>Yang, Xianghui</creatorcontrib><creatorcontrib>Liu, Jian</creatorcontrib><creatorcontrib>Lai, Zeqiang</creatorcontrib><creatorcontrib>Chen, Zhuo</creatorcontrib><creatorcontrib>Liu, Yuhong</creatorcontrib><creatorcontrib>Jiang, Jie</creatorcontrib><creatorcontrib>Guo, Chunchao</creatorcontrib><creatorcontrib>Zhang, Tong</creatorcontrib><creatorcontrib>Gao, Shenghua</creatorcontrib><creatorcontrib>Chen, C. L. Philip</creatorcontrib><title>Scaling Mesh Generation via Compressive Tokenization</title><description>We propose a compressive yet effective mesh representation, Blocked and
Patchified Tokenization (BPT), facilitating the generation of meshes exceeding
8k faces. BPT compresses mesh sequences by employing block-wise indexing and
patch aggregation, reducing their length by approximately 75\% compared to the
original sequences. This compression milestone unlocks the potential to utilize
mesh data with significantly more faces, thereby enhancing detail richness and
improving generation robustness. Empowered with the BPT, we have built a
foundation mesh generative model training on scaled mesh data to support
flexible control for point clouds and images. Our model demonstrates the
capability to generate meshes with intricate details and accurate topology,
achieving SoTA performance on mesh generation and reaching the level for direct
product usage.</description><subject>Computer Science - Computer Vision and Pattern Recognition</subject><subject>Computer Science - Graphics</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNpjYJA0NNAzsTA1NdBPLKrILNMzMjE01DMwNzAy5WQwCU5OzMnMS1fwTS3OUHBPzUstSizJzM9TKMtMVHDOzy0oSi0uzixLVQjJz07Ny6wCS_IwsKYl5hSn8kJpbgZ5N9cQZw9dsPnxBUWZuYlFlfEge-LB9hgTVgEAOAoycw</recordid><startdate>20241111</startdate><enddate>20241111</enddate><creator>Weng, Haohan</creator><creator>Zhao, Zibo</creator><creator>Lei, Biwen</creator><creator>Yang, Xianghui</creator><creator>Liu, Jian</creator><creator>Lai, Zeqiang</creator><creator>Chen, Zhuo</creator><creator>Liu, Yuhong</creator><creator>Jiang, Jie</creator><creator>Guo, Chunchao</creator><creator>Zhang, Tong</creator><creator>Gao, Shenghua</creator><creator>Chen, C. L. Philip</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20241111</creationdate><title>Scaling Mesh Generation via Compressive Tokenization</title><author>Weng, Haohan ; Zhao, Zibo ; Lei, Biwen ; Yang, Xianghui ; Liu, Jian ; Lai, Zeqiang ; Chen, Zhuo ; Liu, Yuhong ; Jiang, Jie ; Guo, Chunchao ; Zhang, Tong ; Gao, Shenghua ; Chen, C. L. Philip</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-arxiv_primary_2411_070253</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Computer Vision and Pattern Recognition</topic><topic>Computer Science - Graphics</topic><toplevel>online_resources</toplevel><creatorcontrib>Weng, Haohan</creatorcontrib><creatorcontrib>Zhao, Zibo</creatorcontrib><creatorcontrib>Lei, Biwen</creatorcontrib><creatorcontrib>Yang, Xianghui</creatorcontrib><creatorcontrib>Liu, Jian</creatorcontrib><creatorcontrib>Lai, Zeqiang</creatorcontrib><creatorcontrib>Chen, Zhuo</creatorcontrib><creatorcontrib>Liu, Yuhong</creatorcontrib><creatorcontrib>Jiang, Jie</creatorcontrib><creatorcontrib>Guo, Chunchao</creatorcontrib><creatorcontrib>Zhang, Tong</creatorcontrib><creatorcontrib>Gao, Shenghua</creatorcontrib><creatorcontrib>Chen, C. L. Philip</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Weng, Haohan</au><au>Zhao, Zibo</au><au>Lei, Biwen</au><au>Yang, Xianghui</au><au>Liu, Jian</au><au>Lai, Zeqiang</au><au>Chen, Zhuo</au><au>Liu, Yuhong</au><au>Jiang, Jie</au><au>Guo, Chunchao</au><au>Zhang, Tong</au><au>Gao, Shenghua</au><au>Chen, C. L. Philip</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Scaling Mesh Generation via Compressive Tokenization</atitle><date>2024-11-11</date><risdate>2024</risdate><abstract>We propose a compressive yet effective mesh representation, Blocked and
Patchified Tokenization (BPT), facilitating the generation of meshes exceeding
8k faces. BPT compresses mesh sequences by employing block-wise indexing and
patch aggregation, reducing their length by approximately 75\% compared to the
original sequences. This compression milestone unlocks the potential to utilize
mesh data with significantly more faces, thereby enhancing detail richness and
improving generation robustness. Empowered with the BPT, we have built a
foundation mesh generative model training on scaled mesh data to support
flexible control for point clouds and images. Our model demonstrates the
capability to generate meshes with intricate details and accurate topology,
achieving SoTA performance on mesh generation and reaching the level for direct
product usage.</abstract><doi>10.48550/arxiv.2411.07025</doi><oa>free_for_read</oa></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | DOI: 10.48550/arxiv.2411.07025 |
| ispartof | |
| issn | |
| language | eng |
| recordid | cdi_arxiv_primary_2411_07025 |
| source | arXiv.org |
| subjects | Computer Science - Computer Vision and Pattern Recognition Computer Science - Graphics |
| title | Scaling Mesh Generation via Compressive Tokenization |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-22T03%3A38%3A29IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Scaling%20Mesh%20Generation%20via%20Compressive%20Tokenization&rft.au=Weng,%20Haohan&rft.date=2024-11-11&rft_id=info:doi/10.48550/arxiv.2411.07025&rft_dat=%3Carxiv_GOX%3E2411_07025%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |