On May 21, 2025, the "2025 STCON Smart Technology and Applications Conference," hosted by ACT International, grandly opened in Suzhou. As a leading enterprise in the "AI + Industrial Vision" field, Mr. Han Xu, the founder of Dongsheng, was invited to attend the conference and delivered a speech titled "Empowering Industrial Vision with AI Large Models: Ushering in a New Era of Intelligent Manufacturing." In his speech, he provided an in-depth analysis of the company's innovative practices and cutting-edge achievements in driving industrial manufacturing upgrades through AI technology.

Mr. Han Xu, CEO of Dongsheng

    To accelerate technological integration and cross-disciplinary collaboration, ACT International, a renowned media group in the high-tech industry with a proven track record of hosting large-scale conferences attended by thousands of participants, has organized a series of events including VisionCon (Visual System Design Conference), LaserFocusCon (Laser Technology Summit), AIMCon (Industrial Technology and Applications Symposium), CRCon (Cleanroom Technology Summit), SBS (Frontier Seminar on New Technologies), CHIP China (Semiconductor Technology Summit), and CSCon (Compound Semiconductor Advanced Technology and Applications Conference). These events are all convening in Suzhou, a pivotal manufacturing hub in the Yangtze River Delta, forming an interconnected technical exchange matrix of multiple forums.

    The gatherings have attracted a multitude of industry experts, scholars, and outstanding enterprises from both domestic and international spheres, fostering in-depth discussions on the application and development of cutting-edge technologies such as machine vision, laser technology, and artificial intelligence (AI) in sectors including semiconductors and electronics manufacturing.

    Leveraging its self-developed AI algorithms and big data analytics technologies, Dongsheng has successfully developed a suite of intelligent industrial AI vision solutions that comprehensively cover multiple critical stages of industrial production, including incoming material inspection, in-process production checks, intelligent production scheduling and management, and final appearance inspection of finished products. These solutions act as the "intelligent engines" of industrial production, effectively enhancing production efficiency and product quality while simultaneously reducing production costs and resource consumption, delivering substantial benefits to industrial manufacturing enterprises.

    In response to the complex, diverse, and highly fragmented quality inspection scenarios in the industrial sector, Dongsheng has constructed an industrial quality inspection large model. This model is built upon vast amounts of heterogeneous product defect data and full-process production information from manufacturing lines, integrating business knowledge in image recognition, defect detection, and process optimization. The model demonstrates strong generalization capabilities, achieving technological breakthroughs in efficient training and optimization, as well as parameter-efficient fine-tuning, thereby reducing training costs and improving development efficiency. Based on this, the Dongsheng Quality Inspection Large Model can be applied across multiple sectors, including new energy lithium batteries, automotive manufacturing, semiconductors, 3C electronics, and pharmaceuticals. It enables functions such as appearance defect detection, real-time production process anomaly alerts, quality risk assessment, process optimization recommendations, and automated generation of quality inspection reports, providing a scientific basis for production management decisions and facilitating the intelligent transformation and upgrading of enterprises.

    Mr. Han Xu shared successful cases of the practical application of Dongsheng's AI technology. The company has engaged in in-depth collaborations with numerous well-known manufacturing enterprises, applying AI technology to real-world production scenarios and achieving remarkable economic and social benefits. Currently, in the pan-3C sector, its inspection services cover comprehensive appearance checks and component inspections for consumer electronics such as mobile phones, laptops, industrial control computers, tablets, and watches. In the automotive field, inspection services include welding defect detection, battery case online inspection, door and cover assembly, gap and flushness measurement, OCR recognition, and stamping part appearance inspection. In the pan-semiconductor sector, inspection services encompass chip, surface-mount LED, and capacitor-inductor inspections. In the lithium battery sector, inspection scenarios cover the entire process inspection and final appearance inspection of lithium batteries, including process inspections and final appearance checks for prismatic, cylindrical, and pouch batteries.

    Leveraging its outstanding technological capabilities and solutions, Dongsheng has successfully delivered projects to global leading companies such as Tesla, enterprises within the Apple supply chain, SONY, Samsung, LG, CISCO, and Schaeffler, facilitating the intelligent transformation of over 15,000 production lines.

    As a frontrunner in the industrial AI sector, Dongsheng consistently leads the industry and redefines standards through its robust technological innovation capabilities. Its advanced technical services and AI solutions have been widely adopted across multiple key sectors, including new energy lithium batteries, automotive manufacturing, semiconductors, 3C electronics, and pharmaceuticals, with business operations spanning 20 countries and regions worldwide. The company has provided comprehensive AI vision inspection solutions to numerous industrial enterprises. Looking ahead, Dongsheng will remain committed to an innovation-driven, win-win cooperation philosophy, collaborating with all stakeholders to advance the deep integration of AI large model technologies in the industrial sector. This effort will support China's manufacturing industry in achieving high-quality development and inject new momentum into the global industrial intelligence landscape.