Group 1: IPO Updates - Hunan Mingming Hen Mang Commercial Chain Co., Ltd. plans to list on the Hong Kong Stock Exchange on January 28, 2026, under the stock code "1768," aiming to become the "first stock of bulk snacks" in Hong Kong [3] - The company intends to issue 14,101,100 H-shares, with a fundraising target between HKD 32.37 billion and HKD 33.36 billion, and a market capitalization between HKD 491.58 billion and HKD 506.56 billion [3] - Mingming Hen Mang has attracted a high-profile cornerstone investor lineup, with eight cornerstone investors subscribing approximately USD 195 million, including Tencent and Temasek [4] Group 2: Company Performance - As of September 30, 2025, Mingming Hen Mang operates 19,517 stores across 28 provinces and all tiered cities in China, becoming the first company in the industry to exceed 20,000 stores [5] - In the first three quarters of 2025, the company achieved a GMV of RMB 66.1 billion, a year-on-year increase of 74.5%, serving 2.1 billion consumers [5] Group 3: Other IPOs - Shanghai Suiruan Technology Co., Ltd. has submitted its IPO application for the Sci-Tech Innovation Board, focusing on AI chip design and development [7] - Suiruan Technology has invested RMB 4.419 billion in R&D from 2022 to the first nine months of 2025, with total revenue reaching RMB 1.654 billion [8] - Shandong Huawutang Cosmetics Co., Ltd. has submitted its IPO application to the Hong Kong Stock Exchange, aiming to become the "first stock of domestic personal care" [10][11] - Shenzhen Wook Feifan Technology Co., Ltd. has also submitted its IPO application, focusing on cross-border retail in Southeast Asia [14][15] Group 4: Financial Highlights - For Huawutang, revenue for 2023 and 2024 is projected at RMB 1.199 billion and RMB 1.499 billion, respectively, with a growth rate of 25% [11] - Wook's revenue for 2023 and 2024 is reported at RMB 908 million and RMB 1.049 billion, with a year-on-year growth of 15.5% [15] - Yunyin Valley Technology Co., Ltd. is recognized as the fifth largest supplier of AMOLED display driver chips globally, with a market share of 40.7% in the Micro-OLED display backplane market [20][21]

证券日报网讯 1月21日，虹软科技在互动平台回答投资者提问时表示，公司采用直销的方式，将计算机 视觉算法技术与客户特定设备深度整合，通过合约的方式授权给客户，允许客户将相关算法软件或软件 包装载在约定型号的智能设备上使用，以此收取技术和软件使用授权费用。公司主要向智能手机、智能 汽车、笔记本电脑、智能家居、智能零售以及各类带摄像头的AIoT设备制造商销售计算机视觉算法软 件及相关解决方案。2025年上半年，公司来源于境外的收入占比为37.28%。 （文章来源：证券日报） ...

智通财经APP获悉，1月21日，中国证监会国际合作司发布《关于山东极视角科技股份有限公司境外发行上市及境内未上市股份"全流通"备案通知书》。公 司拟发行不超过2006.34万股境外上市普通股并在香港联合交易所上市。公司31名股东拟将所持合计99,872,436股境内未上市股份转为境外上市股份，并在香 港联合交易所上市流通。 招股书显示，极视角是中国AI计算机视觉解决方案提供商，为各行各业的企业提供端到端解决方案开发、部署及管理服务。根据弗若斯特沙利文的资料， 按2024年的收入计，公司于中国新兴企业级计算机视觉解决方案市场中排名第八。 附件："全流通"股东名称及转换数量 | 序号 | 股东名称 | 申请全流通股数(股) | | --- | --- | --- | | 1 | 陈振杰 | 16,114,821 | | 2 | 珠海横琴极力投资合伙企业(有限合伙) | 9,452,122 | | 3 | 珠海横琴极创投资合伙企业(有限合伙) | 9,024,164 | | 4 | 深圳市创兴前沿技术股权投资基金合伙企业(有限合伙) | 6.455.286 | | 5 | 青岛经济技术开发区金融投资集团有限公司 | ...

极视角为工业、能源、零售及交通等多个垂直领域的客户提供服务，主要客户来自中国，包括企业、政 府及大学。 中国企业级计算机视觉解决方案市场规模预计将从2024年的人民币368亿元增至2029年的人民币1824亿 元，复合年增长率为37.7%。中国新兴企业级计算机视觉解决方案市场规模预计将从2024年的人民币 111亿元增至2029年的人民币970亿元，复合年增长率为54.3%，其在整体市场中的渗透率将提升至 53.2%。 极视角向港交所主板提交上市申请，中信证券担任独家保荐人。 根据弗若斯特沙利文的报告，按2024年收入计算，极视角在中国新兴企业级计算机视觉解决方案市场中 排名第八，市场份额为1.6%。同时，按相同标准，该公司是中国新兴企业级计算机视觉解决方案市场 第三大以软件为中心的提供商。公司专注于提供AI计算机视觉解决方案(包括标准、定制及软件定义的 一体化AI解决方案)以及大模型解决方案，后者结合客户行业知识，利用多智能体优化、RAG及基于场 景的算法等先进技术，为企业提供定制化服务。 ...

Core Insights - The launch of the PSAI Enterprise Edition by Hongsoft Technology marks a significant advancement in the computer vision industry, transitioning from "point efficiency" to "systematic empowerment" in e-commerce visual production [1][12]. Group 1: Product Features - PSAI Enterprise Edition is designed as a comprehensive AI-driven platform that integrates team collaboration, scalable production, and visual derivative creation for the apparel e-commerce sector [1][4]. - The platform addresses the inefficiencies of traditional visual design processes, which are often cumbersome and costly, by providing a systematic solution that redefines e-commerce visual production [3][4]. Group 2: Production Efficiency - PSAI Enterprise Edition creates a seamless workflow from "raw materials" to "marketing hits," establishing a collaborative and intelligent "e-commerce visual production line" [4][5]. - The platform allows for "one input, multiple outputs," enabling the generation of various visual materials without the need for switching between tools or platforms, thus maximizing efficiency [5][8]. Group 3: Market Adaptation - The platform includes intelligent modules that address complex marketing design needs, such as promotional scenarios and platform-specific requirements, ensuring brand visual consistency across different channels [7][8]. - PSAI Enterprise Edition has been validated in the market, having generated tens of millions of visual materials and partnered with over 1,000 key accounts, demonstrating its reliability and competitive pricing [12][14]. Group 4: Organizational Impact - The platform's dual-drive model of "extreme computing power + agile collaboration" helps brands transition from manual processes to digital factories, significantly enhancing operational efficiency [8][10]. - PSAI Enterprise Edition integrates tools that can drastically reduce production cycles from 20-30 working days to just 1-2 days, with costs dropping by over 95% [8][12].

Core Insights - The article discusses the rapid advancements in 3D Generative Synthesis (3DGS) technology, highlighting its applications in various fields such as 3D modeling, virtual reality, and autonomous driving simulation [2][4] - A comprehensive learning roadmap for 3DGS has been developed to assist newcomers in mastering both theoretical and practical aspects of the technology [4][6] Group 1: 3DGS Technology Overview - The core goal of new perspective synthesis in machine vision is to create 3D models from images or videos that can be processed by computers, leading to numerous applications [2] - The evolution of 3DGS technology has seen significant improvements, including static reconstruction (3DGS), dynamic reconstruction (4DGS), and surface reconstruction (2DGS) [4] - The introduction of feed-forward 3DGS has addressed the inefficiencies of per-scene optimization methods, making the technology more accessible [4][14] Group 2: Course Structure and Content - The course titled "3DGS Theory and Algorithm Practical Tutorial" covers detailed explanations of 2DGS, 3DGS, and 4DGS, along with important research topics in the field [6] - The course is structured into six chapters, starting from foundational knowledge in computer graphics to advanced topics like feed-forward 3DGS [10][11][14] - Each chapter includes practical assignments and discussions to enhance understanding and application of the concepts learned [10][15] Group 3: Target Audience and Prerequisites - The course is designed for individuals with a background in computer graphics, visual reconstruction, and programming, particularly in Python and PyTorch [19] - Participants are expected to have a GPU with a recommended computing power of 4090 or higher to effectively engage with the course material [19] - The course aims to benefit those seeking internships, campus recruitment, or job opportunities in the field of 3DGS [19]

Fund Research - The market saw a majority of equity indices rise, with the largest increase in the ChiNext Index, which rose by 1.86% during the week from December 1 to December 5, 2025 [2] - Public funds experienced significant inflows, with the first ETF tracking the CSI 300 Quality Index closing its fundraising, and Moer Thread emerging as the biggest winner in the offline allocation results [2] - Various equity fund types performed well, with equity funds averaging a rise of 0.93% and a positive return ratio of 76.27% [3] Company Research: Hongsoft Technology (688088) - Hongsoft Technology specializes in AI visual algorithms, providing algorithm licensing and system solutions, with mobile intelligent terminal visual solutions being the main revenue source [5][6] - The company reported a net profit of 142 million yuan for the first three quarters of 2025, marking a year-on-year increase of 60.51% [6] - The global smartphone shipment reached 923 million units in the first three quarters of 2025, with an AI smartphone penetration rate expected to reach 34% [6] - The smart glasses market saw a shipment of 4.065 million units in the first half of 2025, a 64.2% increase year-on-year, with expectations for the market to exceed 40 million units by 2029 [6] - The automotive sector showed a 12.4% year-on-year increase in sales, with the domestic passenger car market's DMS function penetration rate reaching a historical high of 26.2% in September 2025 [6] - The company’s PSAI product has penetrated multiple e-commerce platforms, serving hundreds of thousands of small businesses and over 300 major apparel brands [7] - Earnings per share (EPS) forecasts for 2025-2027 are 0.63 yuan, 0.85 yuan, and 1.18 yuan, with a 2025 price-to-earnings (PE) ratio of 76.07, which is below the average valuation of comparable companies [7] Industry Research: Light Industry Manufacturing & Textile Apparel - The light industry manufacturing sector outperformed the CSI 300 Index by 0.58 percentage points from December 1 to December 5, 2025, while the textile and apparel sector underperformed by 2.88 percentage points [8][10] - Recent price increases in packaging paper have been noted, with companies like Nine Dragons and others announcing price hikes of 50 yuan per ton [10] - The government is actively engaging in trade discussions with the U.S., which may positively impact export chain companies [10] - The domestic real estate market is under pressure, but recent policies to enhance home purchase subsidies may improve the situation in the medium term [10]

Core Viewpoint - The article discusses the SERES (Semantic-Aware Reconstruction from Sparse Views) method, which addresses the challenges of geometric accuracy, detail restoration, and structural integrity in 3D reconstruction from sparse views, providing a low-cost solution to enhance clarity and completeness of geometry [4][27]. Summary by Sections Introduction to SERES - SERES is developed by a collaborative team from Shanghai Jiao Tong University, the University of Manchester, and the Chinese University of Hong Kong, and has been accepted by IEEE Transactions on Visualization and Computer Graphics [6]. Method Overview - The SERES design focuses on two main lines: semantic matching priors and region-level regularization, integrating these into existing frameworks like NeuS or Neuralangelo without altering the core rendering and implicit surface expressions [8]. Semantic Matching Priors - The method involves extracting stable semantic blocks and geometric primitives from input images, allowing for interactive alignment and aggregation across multiple views, which helps the model recognize corresponding details during training [10][12]. Region-Level Regularization - SERES introduces interpretable region consistency in image space, aligning segmented regions with the model's rendered semantic distribution, which provides strong signals for how shapes should align, effectively reducing noise and improving surface coherence [14][22]. Experimental Results - In sparse multi-view settings, SERES significantly improves reconstruction quality and new view synthesis quality, showing a consistent decrease in geometric error as the number of views increases, indicating stable benefits across varying sparsity levels [17][18]. Conclusion - SERES transforms cross-view semantic consistency and structural region constraints into a low-cost, interpretable, and reusable training prior, making it suitable for integration into current sparse 3D reconstruction workflows, achieving high-fidelity geometry with fewer views [27].

Core Viewpoint - Nanling Technology Co., Ltd. plans to invest 100 million RMB in Zhongke Fangcun (Nanjing) Technology Co., Ltd. to acquire a 12.9521% stake, enhancing its market competitiveness and integrating advanced AI technology into its core business [6][7][22]. Group 1: Investment Overview - The investment consists of 30 million RMB for acquiring 4.9345% equity from Nanjing Qilin Venture Capital Co., Ltd. and 70 million RMB for capital increase to obtain an additional 8.4338% equity [9][12]. - The transaction has been approved by the company's board and does not require shareholder approval, as it falls within the board's authority [8][9]. Group 2: Target Company Information - The target company, Zhongke Fangcun, specializes in computer vision AI technology, providing intelligent operation solutions for industries like electric power and petrochemicals [22]. - The company has a registered capital of 11.775413 million RMB and focuses on technology services, software development, and integrated circuit design [11]. Group 3: Strategic Importance - The investment aims to integrate cutting-edge AI technology into Nanling's "Lingyun Service," enhancing its competitive edge in the "cloud-intelligent network security" service matrix [22][23]. - This strategic move is expected to improve the company's overall business resilience and sustainable competitiveness [23]. Group 4: Financial Aspects - The total investment of 100 million RMB is sourced from the company's own or self-raised funds, ensuring no adverse impact on daily operations or cash flow [23]. - The valuation for the equity acquisition aligns with the latest financing round of the target company, indicating a reasonable investment decision [12][22].

Core Insights - The article discusses the development of DA, a novel end-to-end panoramic depth estimator by Tencent's Mixed Reality 3D team, which addresses the challenges of panoramic data scarcity and zero-shot generalization capabilities [2][8]. Group 1: Background and Challenges - Panoramic images provide a 360°×180° immersive view, essential for advanced applications like AR/VR and 3D scene reconstruction [5][6]. - Traditional methods for depth estimation in panoramic images are limited due to the scarcity of panoramic depth data and the inherent spherical distortion of panoramic images [10][12]. - The team aims to expand panoramic data and build a robust data foundation for DA [8]. Group 2: Data Augmentation Engine - The team developed a data management engine to convert high-quality perspective depth data into panoramic data, significantly increasing the quantity and diversity of panoramic samples [11][14]. - Approximately 543K panoramic samples were created, expanding the total sample size from about 63K to approximately 607K, addressing the issue of data scarcity [14]. Group 3: Model Architecture and Training - The SphereViT architecture was introduced to mitigate the effects of spherical distortion, allowing the model to focus on the spherical geometry of panoramic images [16][17]. - The training process incorporates distance loss for global accuracy and normal loss for local surface smoothness, enhancing the model's performance [18]. Group 4: Experimental Results - DA demonstrated state-of-the-art (SOTA) performance, with an average improvement of 38% in AbsRel performance compared to the strongest zero-shot methods [23][24]. - Qualitative comparisons showed that DA's training utilized approximately 21 times more panoramic data than UniK3D, resulting in more accurate geometric predictions [27]. Group 5: Application Scenarios - DA's exceptional zero-shot generalization capabilities enable a wide range of 3D reconstruction applications, such as panoramic multi-view reconstruction [28]. - The model can reconstruct globally aligned 3D point clouds from panoramic images of different rooms in a house or apartment, ensuring spatial consistency across multiple panoramic views [29].