Superb AI Wins 2nd, 4th at International AI Challenge

**Superb AI Dominates CVPR Challenge… Opening a New Horizon for Industrial AI** Superb AI, a specialist in vision AI, achieved brilliant results in the international AI challenge at 'CVPR 2025,' the world's top conference in computer vision, proving the innovative potential of industrial AI technology. Amidst fierce competition, Superb AI secured 2nd place in the 'Individual Object Detection Challenge' and 4th place in the 'Foundation Few-Shot Object Detection Challenge.' Notably, by winning second place, they earned the honor of directly presenting their technology at the CVPR conference, drawing the attention of global AI experts. At the heart of this achievement is 'ZERO,' Superb AI's self-developed vision foundation model. 'ZERO' is a general-purpose AI model applicable to various tasks even without large-scale data training. In this challenge, it showcased an innovative ability to instantly detect objects without the need for data labeling or complex training processes. This represents a significant advancement in 'specific object recognition' technology, essential for solving chronic problems in industrial sites such as defect sorting on manufacturing floors, logistics inventory management, and medical device tracking. Particularly, it holds great significance as it offers a practical solution to reduce the burden of initial costs and specialized personnel for companies hesitant to adopt AI. In the 'Foundation Few-Shot Object Detection Challenge,' 'ZERO' accurately recognized new objects with just 10 example images and text descriptions, successfully demonstrating the applicability of general-purpose AI to complex and diversified video data in various industries such as medical, manufacturing, and agriculture. Recently, as big tech companies concentrate on generative AI and LLM development, Superb AI employed a unique strategy to tackle the challenge of securing high-quality image and video data, which is a difficult problem in creating industrial vision foundation models. Focusing on efficient data construction, lightweight algorithm design, and industry specialization rather than acquiring large-scale data, they demonstrated the prowess to rank among the world's top contenders, surpassing leading universities and research institutions in China. Building on this achievement, Superb AI plans to accelerate the development of practical AI solutions that can be immediately utilized on site in key industrial sectors such as manufacturing, mobility, and physical security, thereby solidifying its leadership in the industrial AI market. Kim Hyun-soo, CEO of Superb AI, emphasized, "This CVPR Challenge achievement proves that we can secure competitiveness that meets global standards even without high-cost, large-scale infrastructure. It will be an important milestone that sets the direction for AI technology to be actually utilized in industrial sites in the future."