科学研究成果
为国家的核心关键技术突破，提供真正中国原创的核心光学、AI技术，布局全球的战略格局。

课题研究


亮亮视野始终坚持核心技术的自主突破和掌握，投入大规模的力量进行技术研发，促成更多具有创新性和影响力的AR+AI技术成果诞生，推动行业的良性快速发展，服务好更多的行业客户。

  • <中关村前沿技术企业任务书
  • 超表面单片三色光栅波导AR光学模组研制>
中关村前沿技术企业任务书
2019年1月-2020年12月
攻关Depthwise卷积算子在Movidius VPU上的优化实现，对比Intel现有的基准实现，2020年优化后的实现运行速度提升5倍，攻关可用于前端设备去重的轻量人脸识别技术。基于薄膜阵列光波导的LEION2/3分体式双目智能眼镜，主要研究在于光学部分的提升。AR智能眼镜是一项具有重大市场潜力的新兴技术，其目标是将真实世界和虚拟世界无缝地结合在一起，最终能够去除显示屏、鼠标和键盘，并可把用户界面变得自然且几乎无法察觉。通过使物理世界更接近数字世界，增强现实必定使二者都具备更高的效率。

超表面单片三色光栅波导AR光学模组研制
2020年7月-2022年7月北京市科委重点研发计划,前沿新材料研究专项课题
针对AR产品对轻小型、隐私性、高能效比、长时间佩戴的需求，开展基于超表面光栅波导的计算全息近眼显示技术研究，突破超表面光栅波导光学系统设计、宽光谱超表面结构设计、计算全息、高精度微纳结构制备等关键设计和加工技术，有效解决结构复杂、体积重量大、角度和波长敏感性强，多级衍射能量损失大、视觉辐辏冲突等不足。

学术合作


亮亮视野高度重视产学研协同创新工作，携手高等院校和科研院所，加强产学研用深度融合，开展核心技术联合攻关，共同致力前沿技术突破，推动学术研究和行业应用结合，共同打造的产学研协同创新生态。

  • <浙江大学光电学院
  • 中科院自动化研究所
  • 清华大学经管学院>
超过8年与浙大光电学院深度合作，亮亮视野依托于浙江大学现代光学仪器国家重点实验室在新型显示技术领域的优势，建立起技术积累、场景驱动研发、工程产品验证的产学研机制。

与中科院自动化所成立了第一视角计算联合实验室，聚焦第一视角计算，并服务于公司安防和工业两大应用场景。具备完善的人脸算法体系，并在这一体系的基础上构建了覆盖云边端安防服务体系。
亮亮视野与清华大学经管学院开展合作，依托“灯塔”项目，深入研究产品的商业价值，通过科学的市场调研分析方法，实现量化定义AR交互范式和行业标准。

专利布局


已申请发明专利近79项，实用新型专利41项，外观设计专利 43项，国际专利2项，软著172项 主要涵盖：光波导光学结构和设计、加工，波导膜系设计，加工和检测方法，决定波导核心指标性能工艺方法，布局未来波导光学技术，虹膜识别，眼动控制交互，AR/VR双模式显示，光场显示。

*部分专利及软著展示

论文成果


以光学显示和AI算法技术体系为根基，布局多领域、多方向前沿研究，目前光学技术在 Optics Letters、Optics Communications 等国际光学期刊上就光波导、OPU光学卷积网络计算单元等科研成果发表文章。AI学术论文国际会刊收录20余篇。

  • <国际光学期刊布局
  • AI 学术论文国际会刊连年收录>
国际光学期刊布局
1. 2021. "Metalens Eyepiece for 3D Holographic Near-Eye Display" Nanomaterials 11, no. 8: 1920.
2. Layered holographic three-dimensional projection using region-partitioned processing,Optics Communications,Volume 481,2021,126553,
3. "Light-field-depth-estimation network based on epipolar geometry and image segmentation," J. Opt. Soc. Am. A 37, 1236-1243 (2020)
4. Tunable beam splitter using bilayer geometric metasurfaces in the visible spectrum. Optics Express. 2020 Sep;28(19):28672-85.
5. Holographic three-dimensional display based on optimizing arrangement of holograms,Optics Communications,Volume 461,2020：125260
6. Designing compact and ultra-efficient illumination lenses with prescribed irradiance properties for extended light sources. Optics Communications. 2020 Jun;465.
7. Precise light control in highly tilted geometry by freeform illumination optics. Optics Letters. 2019 Jun;44(11):2887-90.
8. Development of energy correction based multishot snapshot spectral imaging system. Optics Communications. 2019 Feb;432:116-22.

9." Tunable beam splitter using bilayer geometric metasurfaces in the visible spectrum," Optics Express 28(19), 28672-28685 (2020).

10.Continuous-zoom bifocal metalens by mutual motion of cascaded bilayer metasurfaces in the visible." Optics Express 29.17 (2021): 26569-26585.

11. Metalens Eyepiece for 3D Holographic Near-Eye Display，Nanomaterials 2021, 11(8), 1920（IF 5.1）

12.Helicity-multiplexed bifocal geometric metalenses at visible wavelength. Optical Design and Testing X. International Society for Optics and Photonics, 2020, 11548: 115480A.

13.Transformers in computational visual media: A survey. Computational Visual Media 8(1): 32-62 (2021)

14.Multimodal Global Relation Knowledge Distillation for Egocentric Action Anticipation. ACM Multimedia 2021: 245-254.（CCF-A）

15.Distribution Aligned Multimodal and Multi-Domain Image Stylization. ACM Transactions on Multimedia Computing, Communications, and Applications 17(3): 96:1-96:17 (2021)

16.Unveiling the Potential of Structure-preserving for Weakly Supervised Object Localization. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2021: 11642-11651（CCF-A）

17.Text Style Transfer with Decorative Elements. The 3rd IEEE Workshop on Artificial Intelligence for Art Creation (AIART) 2021

18.Arbitrary Video Style Transfer via Multi-Channel Correlation. The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI) 2021: 35(2), 1210-1217（CCF-A）

19.Visual Tracking with the Siamese Center-Prediction Network. Computational Visual Media 7(2): 253-265 (2021)

20.Destylization of Text with Decorative Elements. ACM Multimedia Asia 2020: 14:1-14:7

21.Arbitrary Style Transfer via Multi-Adaptation Network. ACM Multimedia 2020: 2719-2727（CCF-A）

AI 学术论文国际会刊连年收录
1. Dynamic Refinement Network for Oriented and Densely Packed Object Detection, CVPR 2020 Oral
2. Arbitrary Style Transfer via Multi-Adaptation Network, ACM Multimedia 2020, ACM MM 2020
3. Self-supervised Feature Augmentation for Large Image Object Detection. IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 29, 2020
4. LGM-Net: Learning to Generate Matching Networks for Few Shot Learning, ICML 2019
5. Adaptive Feature Fusion via Graph Neural Network for Person Re-identification, ACM MM 2019
6. tyle-Oriented Representative Paintings Selection, Siggraph Asia 2017
7. Distribution Aligned Multimodal and Multi-Domain Image Stylization. ACM Transactions on Multimedia Computing, Communications, and Applications 17(3): 96:1-96:17 (2021)
8. Unveiling the Potential of Structure-preserving for Weakly Supervised Object Localization. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2021: 11642-11651
9. Text Style Transfer with Decorative Elements. The 3rd IEEE Workshop on Artificial Intelligence for Art Creation (AIART) 2021
10. Arbitrary Video Style Transfer via Multi-Channel Correlation. The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI) 2021: 35(2), 1210-1217
11. Visual Tracking with the Siamese Center-Prediction Network. Computational Visual Media 7(2): 253-265 (2021)
12. Destylization of Text with Decorative Elements. ACM Multimedia Asia 2020: 14:1-14:7
13. Arbitrary Style Transfer via Multi-Adaptation Network. ACM Multimedia 2020: 2719-2727
14. Oilver Deussen, Changsheng Xu: Self-Supervised Feature Augmentation for Large Image Object Detection. IEEE Transactions on Image Processing 29: 6745-6758 (2020)
15. Dynamic Refinement Network for Oriented and Densely Packed Object Detection. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2020: 11204-11213 (Oral)
16. Multi-Attribute Guided Painting Generation. The 2nd IEEE Workshop on Artificial Intelligence for Art Creation (AIART) 2020: 400-403
17. Adaptive Feature Fusion via Graph Neural Network for Person Re-identification. ACM Multimedia 2019: 2115-2123
18. Style-oriented representative paintings selection. SIGGRAPH ASIA (Posters) 2017: 12:1-12:2