PAM团队 投稿 量子位 | 公众号 QbitAI 可以输出语义的「分割一切模型2.0」来了！ 一次交互，「分割+识别+解释+描述」全搞定，同时支持图像、视频和长视频，文本&Mask同时输出！ 由港中文MMLab、港理工、北京大学等机构开源的 PAM （Perceive Anything Model）模型，能够在保留SAM2分割一切、追踪一切能力的 基础上，同时输出丰富的语义信息。 为了训练这样一个强大的模型，PAM团队还构建了一个超大规模高质量训练数据集：拥有 150万个图像区域+60万个视频区域标注 实验结果表明，PAM仅使用 3B参数 ，就在多个图像和视频理解基准上全面刷新或逼近SOTA，且具备更优的推理效率和显存占用，真正实现 性能与轻量的统一。 所有数据均已 完全开源 。 PAM：准确定位一键输出 SAM2拥有强大的分割能力，可以"分割一切物体"，在视频中能够高效追踪 任意目标，表现惊艳！ 但它也有一个明显的局限：无法提供定位目标的任何 语义信息 （比如物体是什么、有何功能、处于什么状态等）。 一些最新的 Video LLM 模型尝试结合VLM和SAM2的强大视觉提示能力，进行视频理解。然而： 这些 ...

在智能手机等显示设备持续追求更优良视觉体验的过程中，中信证券看好LIPO封装工艺的发展趋势， 看好抢先布局、完成产线/客户拉通的相关厂商未来的发展潜力。 ...

机器之心报道 机器之心编辑部 刚刚，在美国田纳西州纳什维尔举办的 CVPR 2025 公布了最佳论文等奖项。 今年共有 14 篇论文入围最佳论文评选，最终 5 篇论文摘得奖项 ，包括 1 篇最佳论文 、 4 篇最佳论文荣誉提名 。此外，大会还颁发了 1 篇最佳学生论文 、 1 篇最 佳学生论文荣誉提名 。 根据会方统计，今年大会共收到 4 万多名作者提交的 13008 份论文。相比去年（11532），今年的投稿数量增长了 13%，最终有 2872 篇论文被接收，整体接收率 约为 22.1%。在接收论文中，Oral 的数量是 96（3.3%），Highlights 的数量是 387（13.7%）。 计算机视觉技术的火热给大会审稿带来了空前的压力。本届投稿作者数量、论文评审者和领域主席（AC）数量均创下新高。 今年前来现场参会的学者也超过 9000 人，他们来自 70 余个国家和地区。 CVPR 官方公布了各个细分领域的论文接收情况，如下图所示。可以看到，图像与视频生成领域今年度的论文接收数量最多，而接收率最高的领域则是基于多视角 和传感器的 3D 以及基于单图像的 3D。 此次，最佳论文奖委员会成员中有 AI ...

Core Insights - Yuanrong Qixing announced its VLA model will be launched to consumers in Q3 2025, with five vehicle models expected to be on the road within the year [1] - The VLA model features four key capabilities: blind spot detection, obstacle recognition, road sign interpretation, and voice control, generating significant interest in the industry [1][3] Company Overview - Yuanrong Qixing, established in 2018 and based in Shenzhen, has focused on autonomous driving and vehicle networking technologies [3] - The VLA model, or Vision Language Action Model, is considered the company's "secret weapon" and offers a differentiating factor compared to traditional end-to-end models by addressing the "black box problem" [3][4] Technology and Innovation - The VLA model enhances transparency by clearly displaying the reasoning process behind its decisions, which increases user trust in the autonomous driving system [4] - In Q4 2024, Yuanrong Qixing captured over 15% market share in the high-level intelligent driving assistance sector with a single mass-produced model [6] - The company has optimized costs through collaboration with Qualcomm, achieving complex scenario operations on a 100 TOPS platform, significantly reducing the price of its intelligent driving solutions [7] Market Challenges - The intelligent driving sector is highly competitive, with many players already established and partnerships formed with automotive manufacturers [8] - Yuanrong Qixing faces challenges in gaining market recognition and acceptance for the VLA model amidst increasing consumer caution due to recent accidents and stringent regulations [8] - The company has successfully raised $100 million in its C1 financing round in November 2024, but still faces financial pressures in a cooling investment environment [8][9] Strategic Considerations - The push to market the VLA model represents both a technological showcase and a market challenge, as the company shifts focus from L4 to L2 capabilities, potentially sacrificing some advanced technology for mass production [9] - The need for ongoing funding is critical to avoid disruptions in technology development and to maintain competitive positioning in the market [9]

2025 年 6 月 13 日 陈伟光 S0340520060001 电话：0769-22119430 邮箱： chenweiguang@dgzq.com.cn 罗炜斌 S0340521020001 电话：0769-22110619 邮箱： luoweibin@dgzq.com.cn 资料来源：iFind，东莞证券研究所 电 子 行 业 双 周 报 需求仍将旺盛 超配（维持） 电子行业双周报（2025/05/30-2025/06/12） 行 业 台系 ODM 厂月度营收同比高增，AI 服务器出货动能充足 投资要点： 本报告的风险等级为中高风险。 本报告的信息均来自已公开信息，关于信息的准确性与完整性，建议投资者谨慎判断，据此入市，风险自担。 请务必阅读末页声明。 电子行业 SAC 执业证书编号： 行情回顾及估值：申万电子板块近2周（05/30-06/12）累计上涨0.67%， 跑输沪深300指数0.20个百分点，在申万行业中排名第16名；板块6月累 计上涨2.57%，跑赢沪深300指数1.21个百分点，在申万行业中排名第14 名；板块今年累计下跌2.01%，跑输沪深300指数0.92个百分点，在申万 行业中排 ...

Core Viewpoint - The application for the IPO of Easy Vision (Hangzhou) Technology Co., Ltd. has been accepted by the Shanghai Stock Exchange, with Guotou Securities as the sponsor [1] Company Overview - Easy Vision focuses on the research, production, and sales of machine vision equipment for automotive manufacturing, providing solutions for various manufacturing processes [1] - The company has developed over ten products that are widely applied in six major manufacturing processes, becoming a supplier for most domestic automotive manufacturers [1] Financial Performance - The company's revenue for the reporting period was 223 million, 355 million, and 392 million yuan, with net profits of 5.39 million, 57.75 million, and 84.43 million yuan respectively [1] - The product categories include visual inspection systems, visual guidance systems, and visual measurement systems [2] Market Position - In 2024, Easy Vision achieved a market share of 22.5% in China's automotive machine vision products, surpassing foreign competitors and becoming the only Chinese company with annual revenue exceeding 100 million yuan in this field [4] Client Network - Easy Vision has established a large and high-quality customer network, with products used by major joint ventures and domestic brands, as well as international companies like Volvo and Rivian [5] - The company has attracted investments from notable institutions, including Guotou Fund and Yinxing Valley Capital [5] Ownership Structure - Before the IPO, the founder, Guo Yin, directly held 12.27% of the company and controlled a total of 56.13% of the shares through various entities [6] IPO Financing - The company plans to raise 1.214 billion yuan through the IPO, which will be used for the industrialization of machine vision products, a research and development center, and to supplement working capital [6]

以下文章来源于巨潮WAVE ，作者小卢鱼 现在的AI应用就像韩国偶像团体一样新人辈出——脸还没被认熟就推出了新的版本，然后就是噼里啪 啦一顿造势，结果总是会有作品、实力配不上流量的感觉。 来源丨巨潮商业评论（ID： tide-biz ） 作者丨 小卢鱼 编辑丨 杨旭然 图源丨Midjourney 而那些十年前出道的AI前辈们，则是大部分都面临着技术路线风险、转型困难、金主不再力捧的问 题，落下了一个或者几个梯队。 这两类企业可以有一个共同的名字：AI掉队者联盟。 以曾经行业头部企业商汤为例，2021年赴港上市的时候，贵为全球估值最高的AI独角兽企业，但从 2022年开始股价就一直深度下跌。目前，其市值与最高点相比已蒸发3000亿港元以上。 其创始团队的退场更是令人唏嘘，科研实力过硬、被称为人脸识别技术开拓者的汤晓鸥先生在2023 年猝然离世，联合创始人徐冰也在今年端午节前官宣辞去执行董事及董事会秘书的职务，获任AI芯片 业务负责人。 商汤在公告中强调，此次职务调整是公司整体人才部署与业务聚焦的一部分，某种程度上也可以理解 为，AI 1.0时代的核心技术并不能用于引领2.0时代。 巨潮WAVE . 融入时代巨潮 ...

Industry Overview - High-speed vision refers to the use of advanced technologies and algorithms that enable machines to analyze and process images at high speeds, facilitating automation and intelligent applications [1][8] - The main products in the high-speed vision sector include high-speed cameras, high-speed vision measurement systems, and related instruments, widely applied in various fields such as industrial inspection, intelligent transportation, and security monitoring [8][11] Global Market - The global high-speed vision market is experiencing accelerated growth, with the market size projected to increase from $8.603 billion in 2017 to $17.954 billion in 2024, and expected to reach $37.489 billion by 2028 [6] Domestic Market - China's high-speed vision industry has seen significant growth, with the market size increasing from 4.735 billion yuan in 2017 to 15.088 billion yuan in 2024, and is forecasted to reach 34.334 billion yuan by 2028 [11][19] Competitive Landscape - The high-speed vision industry in China has developed over the past 20 years, with domestic companies like Zhongke Shijie and Deep Vision emerging as key players, achieving significant market share growth [13][19] - Zhongke Shijie has increased its market share from 8.6% in 2019 to 22.2% in 2022, and in December 2024, it is expected to dominate the market with 58% of the bidding quantity and 49% of the bidding amount globally [13][19] Future Trends - The industry is expected to accelerate technological innovation and domestic substitution, with companies like Zhongke Shijie achieving commercial success with high-frame-rate cameras and other products [19][20] - The application of high-speed vision technology is expanding beyond traditional industrial inspection into sectors like intelligent transportation and medical imaging, with an anticipated annual growth rate of over 20% [20][21] - The industry is moving towards integration and ecosystem building, with domestic companies making breakthroughs in core components and system integration, supported by government and research institutions [21]

激光雷达：AEBS 新规催化标配预期，割草机+无人城配 快速放量 20250612 Q&A 激光雷达在科技行业大规模落地和商业变现中的核心作用是什么？ 科技行业的大规模落地和商业变现的背后，本质上是要解决替代人的问题。替 代人需要模拟人与世界的交互方式，简单来说可以抽象为三个环节：五官感受 客观世界的变化、大脑进行认知和分析、手脚及其他器官反馈和执行。因此， 感知客观世界的变化是所有一切的起点。激光雷达作为多维感知领域的重要技 术之一，在包括智能驾驶、安监等多个方向上发挥了关键作用。 近期车规市场最大的变化是什么？其对激光雷达配置有何影响？ 近期车规市场最大的变化是 ABS 新规征求意见稿的发布。这一新规对于 L2 及 以下级别车辆配置激光雷达，以实现基本安全性防护和感知功能，有非常强的 催化作用。这将推动更多车辆采用激光雷达，提高整体安全性。 激光雷达技术在自动驾驶中的应用呈现下沉化、扩散化、高端化趋势， 价格快速下沉，应用范围从乘用车扩展到轻型商用载货车，同时 L3/L4 级别车辆对高性能激光雷达的需求增加。 无人货运车市场快速发展，多家厂商积极布局，预计 2025 年将实现大 规模交付，激光雷达是主流配 ...

摘要 万马科技通过收购有方科技切入车联网领域，车联网业务收入从 2021 年的 5,000 万元增长到 2024 年的 2.6 亿元，利润也显著提升，并已建 立完整的数据闭环工具链和智驾算力中心。 国内车联网行业渗透率约为 80%，海外市场渗透率不足 30%，随着智 能驾驶对数据需求的增加，国内外市场均有较大的发展空间，尤其 Robotaxi 对实时数据监控和技术要求更高，单车价值提升显著。 优卡科技提供蓝海全球车联和云自动驾驶数据闭环两大解决方案，支持 1,400 万辆车辆，客户包括吉利、上汽、东风和理想等，并在全球范围 内支持 Robotaxi 企业的业务布局。 Robotaxi 被视为车联网行业发展的"皇冠上的明珠"，高盛预测中国 Robotaxi 市场年化增长率将达到 96%。目前已在北京、武汉、广州以 及香港、迪拜等地进行常态化运营，特斯拉也即将推出相关业务。 Robotaxi 运营对网络质量有极高要求，包括运行安全、用户交互、合 规性、自动驾驶数据采集和运维等方面，需要高清地图、车路协同、远 程脱困以及海量数据支持。 万马科技 20250612 据监控需求高，对技术和数据量要求也更高，从单车价值上 ...