J.S. Machine(SZ:000821)2022-12-06 23:36Group 1: Company Overview and Strategic Direction - The company has been focused on packaging machinery since its listing in 1998 and began entering the automotive parts industry in 2008, achieving profitability in recent years [18] - The company has acquired Wuhan Yaohua and Huizhou Sanxie to facilitate strategic transformation and upgrade into precision manufacturing and industrial automation [18] - The focus and resources will be directed towards industrial automation and robotics, with an emphasis on artificial intelligence and core algorithms [19] Group 2: Technological Advancements - Deep learning technology has been successfully applied in the Betel Nut project, enhancing project development and effectiveness [15] - A total of approximately 500 pairs of training-labeled images were prepared for training a 7-layer fully convolutional network, achieving over 99% segmentation accuracy for Betel Nut images [17] - The company plans to develop a convolutional neural network-based operating system to integrate machine vision, embedded systems, and deep learning algorithms for small and medium automation companies [20] Group 3: Market Dynamics and Performance - The packaging machinery market is facing intense competition, with 12 major clients accounting for only about 10% of the market, while in Europe and the US, this figure reaches 50% to 70% [22] - The automotive glass business remains stable, while the automotive casting sector achieved profitability in November last year, with expectations for breakeven or profitability this year [22] - The company aims for a compound growth rate exceeding 30% in the future, focusing on research and development, delivery, and service [21] Group 4: Project Selection Criteria - The company selects projects based on five criteria: labor-intensive nature, potential for standardization, targeting red ocean enterprises, high craftsmanship requirements, and projects that are challenging yet feasible [21] Group 5: Future Outlook - The company recognizes the engineer dividend as a significant advantage for the next decade and plans to continue investing in the robotics industry [19] - The goal is to create a flexible factory capable of personalized production while maintaining consistent costs, leveraging advancements in convolutional neural network technology [20]