近日，慧眼科技与中山大学联合AI算法团队在计算机视觉领域最具影响力的国际顶级学术会议CVPR 2023 Workshop竞赛中，再创佳绩，在Vision-based Industrial Inspection视觉挑战专题1-数据高效缺陷检测的挑战赛中力排国内外强劲对手，获得最佳创新奖，充分展现了慧眼科技在工业AI质检领域的硬核技术实力。

Recently, the AI algorithm team of IE&SYSU made another achievement in the CVPR 2023 Workshop competition, which is the most influential international academic conference in the field of computer vision. In the VISION 23 Challenge Track 1 - Data Efficient Defect Detection Challenge Competition, we beat strong domestic and foreign competitors and won the Best Innovation Award, which fully demonstrates our core technical strength in the field of industrial AI inspection.

CVPR作为计算机视觉和模式识别领域的世界级学术顶会，是计算机视觉领域影响力最深、规模最大的三大国际学术会议之一。作为CVPR 2023的特色项，今年新增以苹果公司为主办方的工业视觉检测专题研习会（1st workshop on Vision-based Industrial Inspection)，特邀苹果供应商、国内外知名高校和科研院所、业内同行共同参与，研习会设立了两项公开挑战赛，有限标注样本条件下的缺陷检测竞赛以及面向缺陷检测的数据生成竞赛，吸引了来自国内外过百组优秀团队同台竞技。慧眼公司受邀参加了有限样本条件下的缺陷检测竞赛，通过激烈角逐，获总榜排名第六，综合技术成果获最佳创新奖的殊荣。

CVPR is one of the three most important international academic conferences with deep and board influence in the field of computer vision. As a new feature of CVPR 2023, 1st workshop on Vision-based Industrial Inspection, organized by Apple, was held this year. Including Apple visual vendors, well-known universities and research institutes, and machine vision peers, were specially invited to participate in this workshop. There are two opening challenge competitions, the defect detection competition under the condition of limited labeled samples and the defect detection-oriented data generation competition, attracting more than 100 outstanding teams from home and abroad to compete on the same stage. IE was invited to participate in the frist competition and finally won the best innovation award for its comprehensive technical achievements.

作为人工智能技术领域的领军者与深耕者，慧眼科技坚持自主研发的发展路线，在工业AI质检技术领域拥有强大的技术优势和深厚应用积累。算法团队则依托深度学习、半监督学习、迁移学习以及自动数据增广技术，使算法在多机台、多点位下对类似缺陷实现稳定检测；与此同时，从算法的冗余性角度出发，结合正样本学习保证了罕见缺陷及异常的零漏检。在此基础上，依托公司标准化的柔性质检平台（灵动仪及灵动线）形成系列一体式的智能解决方案，突破制造业外观检测自动化的瓶颈问题，大幅提升检测效率，解放质检人力、优化成本。

As a pioneer and deep cultivator in the field of artificial intelligence technology, IE adheres to the independent research and development routes, and has strong technical advantages and profound application accumulation in the field of AI industrial visual inspection technology. The algorithm team achieves stable detection of similar defects in multi-machine and multi-position based on deep learning, semi-supervised learning, transfer learning and automatic data augmentation techniques. Meanwhile, positive sample learning is used to avoid the missing detection of rare defects and anomalies . On this basis, combining with the standard flexible visual inspection platform (FlexView AIV and FlexLine), we form a series of integrated intelligent solutions, breaking the bottleneck of automatic appearance inspection, greatly improve detection efficiency, free quality inspection manpower and optimize overall cost.

今后，慧眼科技将切实加强人工智能技术创新研发和智能制造服务，加强创新开拓与落地应用，力争成为机器视觉人工智能产业发展的中坚力量。

In the future, IE will strengthen artificial intelligence technology innovation R&D and intelligent manufacturing services, strengthen innovation development and application, and strive to become the backbone of the development of machine vision artificial intelligence industry.