[Gettyimages Bank]
Data from Statistica, a global market research firm, showed that as of 2018, there are 1.03 million CCTVs in public locations in South Korea. Millions of surveillance cameras are also installed in commercial buildings, shops, houses and other private places. CCTV footage is widely used as a crime prevention method and forensic evidence.
Camera functions are getting more powerful thanks to technological advance, but they are not a perfect solution because of relatively low resolution. The Electronics and Telecommunications Research Institute (ETRI), a state-funded research body, said on Friday that it has developed an artificial intelligence technology that can identify fuzzy vehicle numbers on CCTVs.
The institute said its technology can restore images of blurred or broken pictures taken from various angles and learn them through deep learning methods to derive numbers. A model that learns false data and a model that discovers it is competing with each other, creating data that is closer to reality.
In a contest with 30 people the previous day to identify license plates of vehicles, the institute said its AI scored 82 points, far outpacing a person's highest score of 21 points. Kwon Tae-hyung, a researcher at the Institute for Security Policy at the National Police University, said ETRI's technology contributes to the realization of smart security.
Copyright ⓒ Aju Press All rights reserved.
Related News
-
Unlocking Investment: AI Transition Challenges for Small Businesses
“To transition to artificial intelligence (AX), significant costs are unavoidable, starting from attaching labels and sensors for data accumulation, installing closed-circuit televisions (CCTV), to hiring personnel for data refinement and utilization, and purchasing robots,” said a representative from a manufacturing company in Daegu. This sentiment was echoed in a recent survey conducted by the Main Biz Association, which involved 307 main business companies across the country. According to the association, 71.4% of responding companies acknowledged the necessity of adopting and utilizing AI for securing competitiveness and sustainable growth. However, only 17.6% of these companies are currently developing concrete plans for expanding and advancing AI adoption within the next two years. Additionally, 33.6% of respondents recognized the need for AI but reported having no specific plans in place. During the consideration and preparation stages for AI adoption, the main challenges identified were a lack of suitable AI use cases (49.2%) and uncertainty regarding expected benefits and investment returns (31.6%). In the implementation and validation stages, difficulties in modifying systems to fit specific job characteristics (44.6%) and challenges in integrating existing systems and data (38.5%) were highlighted as significant obstacles. While step-by-step investments are necessary for discovering initial use cases, building systems, and securing operational costs and capabilities, small businesses face a much greater burden compared to large corporations. A survey by the Korea Chamber of Commerce and Industry revealed that 73.6% of responding manufacturing companies feel pressured by AI investment costs, with small businesses (79.7%) expressing more concern than large companies (57.1%). Calls for tailored, step-by-step support for AI adoption are growing, alongside demands for regulatory reforms to facilitate investment in cutting-edge technologies like AX for small businesses. In the new industry sector, regulations surrounding personal data and copyright are seen as major barriers to acquiring the data necessary for AI development. For physical AI applications, such as robotics and drones, capturing video and behavioral data in real environments is essential, yet current regulations limit their utilization. Advocates argue that related systems should be reformed to align with those of leading countries. There have also been ongoing concerns regarding delays in legal adjustments following regulatory sandbox trials. Regulatory sandboxes allow new industries or technologies to test in the market under a temporary suspension of existing regulations. However, if legal adjustments are delayed despite no significant risks arising from trial exceptions, companies may face uncertainty in formalizing their businesses. Won So-yeon, head of the Regulatory Policy Research Office at the Korea Administrative Research Institute, stated, “The Regulatory Rationalization Committee should be empowered to regularly review and adjust the issues related to delays in legal adjustments. There is also a need to clarify the criteria for legal adjustments at the trial approval stage.” Regarding the growth of small businesses, easing regulations on the KOSDAQ market is a key issue. The industry is concerned that delays in listing reviews increase costs and administrative burdens for companies, while criteria for delisting are rapidly tightening, potentially pushing even normally operating small and venture businesses out of the market. There are calls to assess delisting not solely based on market capitalization and stock prices but to consider financial status, technological capability, and growth potential comprehensively. There are also demands to improve the lock-up practices for newly listed KOSDAQ companies. The venture capital sector argues that prolonged voluntary lock-ups beyond mandatory requirements delay the recovery of venture investments and reduce the capacity for new investments, necessitating a reduction in the burden of lock-ups on long-term investments. Furthermore, there are ongoing calls to relax industry restrictions in local industrial complexes and to expand small business participation in government-supported mega-projects, such as large AI data centers. In aging industrial complexes, past restrictions on resident industries do not reflect current industrial structural changes, prompting calls to broaden industry exceptions and to establish incentives for prioritizing small business products in government-supported large projects through procurement and tax benefits.* This article has been translated by AI.
-
CJ CheilJedang Introduces 'Physical AI' in Food Manufacturing for Autonomous Production
CJ CheilJedang is set to develop technology that utilizes artificial intelligence (AI) to automatically manage and control food production processes. The company plans to enhance productivity and quality competitiveness by applying AI throughout the manufacturing site, including real-time detection of product defects and predictive maintenance of equipment.On November 11, CJ CheilJedang announced its participation as a representative company in the food sector for the Ministry of Trade's 'AI Factory Support Project.'The core of this initiative is the development of 'Physical AI' technology that controls actual equipment and processes. CJ CheilJedang will collaborate with WonPredict, Ivex, and the Korea Institute of Industrial Technology to create autonomous manufacturing and integrated predictive maintenance technologies tailored for the food and beverage (F&B) industry.Initially, the company will establish a foundation for real-time data collection and analysis from various equipment in the manufacturing environment. Based on this, it plans to apply 'Physics-Informed Neural Networks (PINN)' technology to learn the physical characteristics of processes and sensor data, predicting optimal equipment operating conditions and automatically controlling production processes.AI will also be utilized in quality management. By combining machine vision that assesses product conditions with robotics and edge AI technology that processes data on-site, the system will detect foreign substances, product damage, and packaging defects in real time. If anomalies are identified, the system will respond by coordinating with equipment control to address defects.In equipment management, AI will analyze the operational status of machinery to detect signs of failure in advance, establishing a predictive maintenance system that minimizes unexpected production line shutdowns.AI CCTV will respond to situations where workers collapse or enter hazardous areas. The technology will develop safety control measures that automatically slow down or stop production equipment upon detecting dangerous situations in real time, thereby preventing industrial accidents.CJ CheilJedang plans to integrate these technologies using the digital transformation (DX) foundation established through its existing smart factory and production automation systems. The goal is to implement an autonomous manufacturing system where AI recognizes the state of the production site, predicts anomalies, and determines necessary actions, including equipment control.Kang Min-soo, Vice President of Production at CJ CheilJedang, stated, "In food manufacturing, the characteristics of raw materials and changes in the production environment directly affect product quality. Therefore, AI technology that understands both on-site data and process characteristics is crucial. We aim to secure Physical AI technology optimized for F&B manufacturing to enhance productivity, quality, and safety competitiveness, while also strengthening the global competitiveness of K-food."* This article has been translated by AI.
-
Alice Huang's Background Raises Suspicion Amid Allegations of Espionage
Questions are mounting regarding Alice Huang, a Chinese businesswoman identified as the alleged mistress of Jensen Huang, CEO of NVIDIA. Her age, educational background, and past career remain unclear, and doubts have also been raised about the business operations of Megaspeed, the Singapore data center company she managed. According to reports from The Guardian and The Wall Street Journal on November 9, Alice Huang previously worked as a journalist for China Central Television (CCTV) and Hunan TV before holding positions at financial firms such as HSBC and China Merchants Bank. Her profile on a Chinese private equity information site claims she served as an assistant to the head of HSBC's private banking division, vice president at China Merchants Bank, and executive director and fund manager at Shanghai Hushi Investment Management. However, much of this career information remains unverified. An executive from Hunan TV stated they did not know Alice Huang, and both HSBC and China Merchants Bank did not respond to media inquiries regarding her employment. Some of her executive role at Shanghai Hushi Investment Management has been confirmed through past company postings. Doubts also surround the legitimacy of Megaspeed, where she was an executive. The Straits Times reported on November 10 that the office of Megaspeed, located in a high-end office building in Singapore's central business district, was closed last October, and phone calls went unanswered. Other tenants in the building noted that the lights in the office were always off, and they rarely saw employees coming and going. A reporter from the South China Morning Post also faced denial of entry when attempting to visit the office recently. Megaspeed's website claims it is a key partner for NVIDIA's cloud computing business in the Asia-Pacific region and asserts that it operates data centers in various Southeast Asian locations. However, its LinkedIn profile lists only 2 to 10 employees. Despite this, it has been reported that through its Malaysian subsidiary, Speedmetrics, Megaspeed imported NVIDIA AI equipment worth at least $4.6 billion. U.S. authorities have been investigating the company since 2024 over suspicions that it may have smuggled advanced semiconductors restricted for export to China. Megaspeed began as a Singapore subsidiary of the Chinese gaming company Seven Roads in 2015 and changed its name in March 2023. At that time, Alice Huang secured ownership and served as managing director but transferred the company to Singaporean Tan Yong Pong in November of the same year and stepped down from management. Currently, the sole registered shareholder of Megaspeed is Swift Data, owned by Tan. However, the registered address of this company is reportedly not an independent office but a location previously used by another professional service firm, raising further questions. As a result, suspicions are growing that Alice Huang may be an agent recruited by Chinese intelligence, as U.S. intelligence agencies suspect she may have been gathering information related to advanced AI semiconductors for Chinese intelligence. However, no evidence has been presented to prove her espionage activities. The Chinese government has denied the allegations, and NVIDIA stated that its internal investigation into Megaspeed found no concerning issues. Singapore police continue to investigate potential legal violations by Megaspeed but clarified that Alice Huang's connections to the Chinese government or intelligence agencies are not part of their inquiry.* This article has been translated by AI.