捷成股份发布四阶影视理解大模型 AI视觉中枢重构产业生态

Core Insights - The 2025 China Digital Entertainment Conference highlighted the transformative potential of the four-tier film and video understanding model developed by Jiecheng Technology, which aims to reconstruct the entire film industry chain [1][4] - The model addresses the industry's core pain points of high understanding costs and low expression efficiency, driven by AIGC technology [1][3] Group 1: Technology and Innovation - Jiecheng Technology's four-tier film video understanding model establishes the industry's first standardized technical evaluation framework, featuring layers that enhance memory retrieval, narrative restructuring, multi-perspective analysis, and creative simulation of directorial styles [2][3] - The model enables a cognitive leap in film understanding, transitioning from basic comprehension to deeper semantic support for content creation [3] Group 2: Industry Impact and Collaborations - The company has built a substantial video corpus, leveraging over 200,000 hours of high-quality audio-visual material, including major works, to support the model's development [3] - Strategic partnerships with leading IP content platforms like Tomato and Hongguo have been established to create new AI-driven content, aiming for a production capacity of 1,000 AI short dramas [4] - The model's commercial application is expected to enhance the global competitiveness of Chinese digital cultural products, positioning Jiecheng Technology as a leader in AI applications within the film industry [4]