En3D vs Pixso AI：怎么选？

关于 En3D

Papersarxiv:2401.01173Copy markdownEn3D: An Enhanced Generative Model for Sculpting 3D Humans from 2D Synthetic DataPublished on Jan 2, 2024·Submitted byAKon Jan 3, 2024Upvote12+4Authors:Yifang Men,Biwen Lei,Yuan Yao,Miaomiao Cui,Zhouhui Lian,Xuansong XieAbstractEn3D generates high-quality 3D human avatars using synthetic 2D data without relying on pre-existing 3D models, employing a combination of a 3D generator, geometry sculptor, and texturing module.Generated byQwen/Qwen2.5-Coder-32B-InstructWe present En3D, an enhancedgenerative schemefor sculpting high-quality 3D human avatars. Unlike previous works that rely on scarce 3D datasets or limited 2D collections with imbalanced viewing angles and imprecise pose priors, our approach aims to develop a zero-shot 3Dgenerative schemecapable of producing visually realistic, geometrically accurate and content-wise diverse 3D humans without relying on pre-existing 3D or 2D assets. To address this challenge, we introduce a meticulously crafted workflow that implements accurate physical modeling to learn the enhanced3D generative modelfrom synthetic 2D data. During inference, we integrate optimization modules to bridge the gap between realistic appearances and coarse 3D shapes. Specifically, En3D comprises three modules: a3D generatorthat accurately models generalizable 3D humans with realistic appearance from synthesized balanced, diverse, and structured human images; ageometry sculptorthat enhances shape quality using multi-view normal constraints for intricate human anatomy; and atexturing modulethat disentangles explicit texture maps with fidelity and editability, leveragingsemantical UV partitioningand adifferentiable rasterizer. Experimental results show that our approach significantly outperforms prior works in terms of image quality, geometry accuracy and content diversity. We also showcase the applicability of our generated avatars for animation and editing, as well as the scalability of our approach for content-style free adaptation.View arXiv pageView PDFAdd to collectionCommunitymiaoyinJan 4, 2024This comment has been hiddenmiaoyinJan 4, 2024This comment has been hiddenAmirsefatJan 19, 2024A man with a womanReplyChromaFlowFeb 10, 2024This comment has been hiddenLucas3467May 18, 2024ReplywwwguruJun 23, 2024This comment has been hiddenwwwguruJun 23, 2024Перерисуй фото в 3DReplyAnDongEluosiOct 11, 2024ReplyAditya98Jan 22, 2025convert this watch into a 3d model whichi is seen by realistic ,See translationReplyEditPreviewUpload images, audio, and videos by dragging in the text input, pasting, orclicking here.Tap or paste here to upload imagesComment·Sign uporlog into comment

AbstractEn3D generates high-quality 3D human avatars using synthetic 2D data without relying on pre-existing 3D models, employing a combination of a 3D generator, geometry sculptor, and texturing module.Generated byQwen/Qwen2.5-Coder-32B-InstructWe present En3D, an enhancedgenerative schemefor sculpting high-quality 3D human avatars. Unlike previous works that rely on scarce 3D datasets or limited 2D collections with imbalanced viewing angles and imprecise pose priors, our approach aims to develop a zero-shot 3Dgenerative schemecapable of producing visually realistic, geometrically accurate and content-wise diverse 3D humans without relying on pre-existing 3D or 2D assets. To address this challenge, we introduce a meticulously crafted workflow that implements accurate physical modeling to learn the enhanced3D generative modelfrom synthetic 2D data. During inference, we integrate optimization modules to bridge the gap between realistic appearances and coarse 3D shapes. Specifically, En3D comprises three modules: a3D generatorthat accurately models generalizable 3D humans with realistic appearance from synthesized balanced, diverse, and structured human images; ageometry sculptorthat enhances shape quality using multi-view normal constraints for intricate human anatomy; and atexturing modulethat disentangles explicit texture maps with fidelity and editability, leveragingsemantical UV partitioningand adifferentiable rasterizer. Experimental results show that our approach significantly outperforms prior works in terms of image quality, geometry accuracy and content diversity. We also showcase the applicability of our generated avatars for animation and editing, as well as the scalability of our approach for content-style free adaptation.

En3D generates high-quality 3D human avatars using synthetic 2D data without relying on pre-existing 3D models, employing a combination of a 3D generator, geometry sculptor, and texturing module.Generated byQwen/Qwen2.5-Coder-32B-InstructWe present En3D, an enhancedgenerative schemefor sculpting high-quality 3D human avatars. Unlike previous works that rely on scarce 3D datasets or limited 2D collections with imbalanced viewing angles and imprecise pose priors, our approach aims to develop a zero-shot 3Dgenerative schemecapable of producing visually realistic, geometrically accurate and content-wise diverse 3D humans without relying on pre-existing 3D or 2D assets. To address this challenge, we introduce a meticulously crafted workflow that implements accurate physical modeling to learn the enhanced3D generative modelfrom synthetic 2D data. During inference, we integrate optimization modules to bridge the gap between realistic appearances and coarse 3D shapes. Specifically, En3D comprises three modules: a3D generatorthat accurately models generalizable 3D humans with realistic appearance from synthesized balanced, diverse, and structured human images; ageometry sculptorthat enhances shape quality using multi-view normal constraints for intricate human anatomy; and atexturing modulethat disentangles explicit texture maps with fidelity and editability, leveragingsemantical UV partitioningand adifferentiable rasterizer. Experimental results show that our approach significantly outperforms prior works in terms of image quality, geometry accuracy and content diversity. We also showcase the applicability of our generated avatars for animation and editing, as well as the scalability of our approach for content-style free adaptation.

核心功能

• Join the discussion on this paper page（自动优化版）
• En3D generates high-quality 3D human avatars using synthetic 2D data without relying on pre-existing 3D models, employing a combination of a 3D generator, geometry sculptor, and texturing module.（自动优化版）
• A man with a woman（自动优化版）
• Перерисуй фото в 3D（自动优化版）
• convert this watch into a 3d model whichi is seen by realistic ,（自动优化版）
• ·Sign uporlog into comment（自动优化版）
• Get this paper in your agent:（自动优化版）
• No dataset linking this paper（自动优化版）

Pixso AI是Pixso推出的一款集成式AI设计工具，专为提升设计效率而打造。它通过人工智能技术简化从创意到成品的全流程。核心功能包括AI生图，能根据文字描述快速生成多风格的设计素材；AI抠图与背景替换，可一键分离主体并智能填充新背景；以及AI设计规范生成，自动分析并创建统一的设计系统。这款工具面向UI/UX设计师、平面设计师以及需要快速产出视觉内容的产品经理和运营人员。使用场景广泛，例如在项目初期快速生成界面原型和图标，在电商设计中批量处理产品图，或是在团队协作中自动维护设计规范，减少重复性工作。