Breaking World Records Repeatedly, CloudWalk Technology Wins Three Firsts in the Global Authority FRVT Test
Recently, the global authority face recognition algorithm test (NIST-FRVT) released its latest list. CloudWalk Technology's research institute won the championship in all three tracks: face recognition 1:1, face recognition 1:N, and masked face recognition. Far ahead of other manufacturers and research institutions such as Toshiba and Samsung, it demonstrated strong technical strength and amazed the world with Chinese AI power.
1:1Test
1:NTest
Mask occlusion test
ChallengeHundreds of billions of comparisons Global manufacturers competeCposition
NIST-FRVT conducts hundreds of billions of comparisons using different types of photo samples, and the evaluation of face recognition algorithms can reach one-millionth accuracy. It is recognized as the world's largest, most stringent, most competitive, and most authoritative face recognition algorithm test, known as the gold standard in the industry. The test set is rich, covering a comprehensive range of scenarios, requiring the model to have sufficient generalization. At the same time, FRVT data comes from real business scenarios, and the test results represent the performance of the technology in actual combat, better testing the implementation value of the algorithm.
This test is a competition arena for the world's top research institutions. The test data is not disclosed, and tests are conducted based on the technology provided by the participating teams to ensure objective fairness. To date, nearly 100 companies and research institutions worldwide have participated in this test, including SAMSUNG (South Korea), TOSHIBA (Japan), VisionLabs (Russia), Ever AI (USA), Vocord (Russia), CloudWalk, Meituan, Hikvision, Dahua, Tencent, etc.
Tackling core key technologies Building ubiquitous intelligent cities
CloudWalk Technology's visual algorithm design takes model generalization and robustness as the primary goal, rather than leaning towards one or more specific test scenarios. The CloudWalk technical team stated that only by adhering to this design principle can value be realized in actual products. In the face of the global pandemic situation, the product not only emphasizes recognition performance under mask occlusion but should also consider the coexistence of no mask and mask, improving the model's generalization ability to cope with more complex real-world situations, thereby truly playing the role of the algorithm.
Over the years, under the leadership of founder Zhou Xi, CloudWalk's AI R&D team has brought together many outstanding talents from home and abroad, possessing independently controllable and continuously innovative full-chain AI technology, achieving a closed loop from intelligent perception to cognition and decision-making.
CloudWalk Technology not only focuses on continuous leadership in algorithm research but also continuously promotes the industrial implementation of AI technology, effectively applying various advanced technologies to various fields, empowering the intelligent transformation of various industries, and transforming technological achievements into intelligent power that serves society and people's lives.
Through continuous technological innovation and solid business development, CloudWalk Technology has achieved mature applications in areas such as Smart Finance, Smart Governance and Smart City, Smart Mobility, and Smart Commerce. Further promoting the deepening of a complete industry chain closed loop based on CloudWalk's Human-Machine Collaborative Operating System, it will provide more comprehensive and efficient human-machine collaborative solutions in the future.
You May Be Interested In
-
2023-10-19CloudWalk's visual foundation large model demonstrates strong generalization performance, greatly reducing the data dependency and development cost required for downstream tasks. At the same time, zero-shot significantly improves training and development efficiency, making widespread application and rapid deployment possible.
-
2022-12-26The globally authoritative NIST Face Recognition Vendor Test (NIST-FRVT) released its latest rankings, with CloudWalk achieving three first-place results.
-
2025-12-12When both the tracker and the target are flying at low altitude and high speed, can traditional visual tracking algorithms keep up? Recently, a joint research team from CloudWalk Technology, Shanghai Jiao Tong University, Hong Kong University of Science and Technology (Guangzhou), Sun Yat-sen University, and the Institute of Information Engineering, Chinese Academy of Sciences released a hardcore work—UAV-Anti-UAV. This is the industry's first million-level multimodal anti-UAV visual tracking benchmark for "Air-to-Air" scenarios, and proposes a powerful Mamba-based baseline, MambaSTS. MambaSTS achieves the best performance on all five metrics of the UAV-Anti-UAV benchmark. This is another technological breakthrough by CloudWalk Technology in multimodal large models. Facing dual dynamic interference, how does the current SOTA perform? Let's find out!
