Core Viewpoint - The article discusses the innovative EyeReal technology, which enables glasses-free 3D display, marking a significant advancement in visual technology and potentially transforming various industries such as gaming, education, and virtual reality [2][6][17]. Group 1: Technology Overview - EyeReal technology allows for a glasses-free 3D display with a viewing angle exceeding 100 degrees, providing a smooth visual experience even when the viewer moves [6][7]. - The effective 3D imaging area of EyeReal is between 0.1 to 0.2 square meters, which is 1000 times larger than previous holographic technologies that could only display images the size of a fingernail [9]. - It achieves true "full parallax" display, supporting horizontal, vertical, and radial viewing, allowing for realistic geometric perspective changes as the viewer moves [10][11]. Group 2: Technical Innovations - EyeReal integrates computational optics with artificial intelligence, utilizing a new strategy called dynamic spatial-bandwidth product (SBP) utilization to overcome physical limitations of traditional 3D displays [17][18]. - The system employs real-time eye-tracking to project light precisely to the viewer's eyes, enhancing the viewing experience by aligning the light field with the viewer's retinal position [19][21]. - A key technology, "ocular geometric encoding," allows for reverse perspective transformation, ensuring that images are accurately projected based on the viewer's eye position [25][26]. Group 3: Hardware and Performance - The hardware consists of a multi-layer panel structure with three layers of TFT-LCD panels, spaced approximately 3 centimeters apart, working with white LED light sources and orthogonal polarizers [30]. - EyeReal maintains a high-definition resolution of 1920×1080 and a refresh rate exceeding 50Hz, ensuring smooth and dynamic content display [15]. - The AI-driven system uses a lightweight fully convolutional neural network (FCN) to modulate optical signals and synthesize images, ensuring high-quality visual output [32][35]. Group 4: Research and Development - The lead author of the research is a 26-year-old PhD student from Fudan University, indicating a strong academic foundation and innovative potential in the field of AI and visual technology [5][44]. - The research team includes notable figures from Shanghai AI Lab and other institutions, highlighting collaboration across academia and industry [47][49].

以下文章来源于硬氪 ，作者张卓倩 硬氪 . 视频实时转化处理生成3D内容，并适配各类型3D显示屏。算法还支持实时动态合成3D数字人和3D场景信息，实现"云合影"，并 可应用于车载3D HUD场景，实现增强型辅助驾驶和3D立体导航效果。 封面来源 ｜ IC photo 36氪获悉，苏州智聚芯联微电子有限公司（以下简称"智聚芯联"）近日完成数千万元Pre-A轮融资，由稳致资本领投。本轮资金将 用于支撑企业研发投入及产线采购，推动裸眼3D显示技术的产业化进程。截止本轮，公司已陆续完成了3轮股权融资。 "智聚芯联"成立于2021年5月，是一家专注于三维空间现实显示技术，从光路设计、到3D空间视觉算法、3D模组适配，提供整个 3D空间视觉解决方案的服务商。公司产品适配LCD、OLED、MiniLED等多种显示屏幕，其技术广泛应用于广告、娱乐、医疗、 美容和智能家居等多个行业。在用户端，"智聚芯联"可以基于用户的显示设备，快速捕捉生成数字人。2023年底公司推出了第一 代原型机。 3D显示行业当前正经历着迅猛的发展，其技术主要分为眼镜式和裸眼式。裸眼3D技术目前主要用于公共商务场合，并有望在未来 应用于便携式设备。根据G ...

与此同时，裸眼 3D 显示行业也面临诸多挑战。当前，裸眼3D显示的产业化主要受制于极高的硬件成 本、较低的分辨率以及高昂的内容创作成本。特别是在广播电视领域，高效经济地实现从2D到3D内容 的即时转换，构成核心挑战。"针对目前行业的两大痛点，我们提出全栈解决方案，主要就是解决裸眼 3D产品落地最后一公里的问题"，「智聚芯联」创始人关宇昕介绍，"我们在成本上的控制，可以让用 户用一杯奶茶钱就可以体验裸眼3D技术，感受不一样的'幻境'世界。" 「智聚芯联」凭借半导体工艺技术，自主研发的光学模组采用纳米压印技术，具有加工精度高、成本 低、适配性强等特点。公司独立研发的「2D - TO - 3D」算法大模型和渲染引擎，能够实现图像的高效 处理和高质量转换。该算法支持环拍相机阵列一秒采集人体 3D 信息，并配合管线构建真实的 3D 数字 人。同时，利用端侧小模型 AIGC 技术，可以通过图像构建逼真的 3D 模型，实现 2D 图像和视频实时 转化处理生成 3D 内容，并适配各类型 3D 显示屏。算法还支持实时动态合成 3D 数字人和 3D 场景信 息，实现「云合影」，并可应用于车载 3D HUD 场景，实现增强型辅助驾 ...