Deeply Showcases Innovative Listen AI Solution at CES 2025 and Intersec
관리자
2025-01-08
Deeply, an AI-powered sound analysis company, unveiled its innovative solution, Listen AI, at CES 2025. This groundbreaking technology offers safety solutions for detecting emergency sounds, such as screams and shouts, in physical security settings, as well as machine solutions capable of identifying abnormal mechanical noises in industrial environments. To support these solutions, Deeply introduced a suite of proprietary tools, including three specialized microphones, the AI Sound Box, and a real-time monitoring dashboard.
CES 2025, the world’s largest IT and electronics trade show, was held from January 7 to 10 in Las Vegas. On the opening day, Deeply demonstrated the capabilities of Listen AI during the Korea Pavilion product showcase, highlighting the potential of its technology to a global audience. Following CES, the company plans to participate in the Intersec exhibition in Dubai to further demonstrate its advanced solutions on the global stage.
Listen to AI addresses critical safety needs by detecting emergency sounds such as screams, shouts, and cries for help in real-time, offering a significant improvement over traditional camera-based surveillance systems. With minimal privacy concerns, the technology is suitable for sensitive areas such as restrooms and changing rooms. The solution is already enhancing public safety in locations such as Incheon Transit Corporation’s subway stations and the sports facilities of the Government Complex in Sejong.
In addition to its safety applications, Listen AI is also transforming industrial maintenance by detecting mechanical issues, including abnormal engine noises, bearing faults, and fastening irregularities. By predicting potential failures and extending the lifespan of machinery, this technology is proving invaluable in sectors such as automotive manufacturing, high-speed KTX trains, and military unmanned drones.
Deeply’s innovations are underpinned by the world’s largest non-verbal sound dataset, comprising 13.9TB of data. Using this resource, the company has developed a transformer-based deep learning model with over 98% accuracy. Leveraging MLOps-driven automated learning systems Deeply has reduced false positives by a factor of 10, ensuring continuous improvement in model performance. These features make the solution highly adaptable to real-world applications, ranging from unmanned monitoring to smart cities and factories. With this focus, Deeply is positioning itself as a leader in auditory AI, particularly in the emerging multimodal AI era.
Through CES, Deeply introduced a fully developed Listen AI solution, designed for immediate global deployment. The company also showcased a revamped monitoring dashboard, which significantly enhances ease of use and accuracy, enabling real-time situational awareness and swift responses. Starting February 1 Deeply plans to offer products through a CES-exclusive promotion with global shipping availability, aiming to forge new partnerships and expand its market presence.
Deeply CEO Suji Lee stated, “Our participation in CES sets a new benchmark for sound AI technology, establishing Deeply as a global leader driving innovation in both manufacturing and safety sectors.” By unveiling Listen AI on the international stage, Deeply is redefining the future of sound-based AI, presenting its vision for safer and smarter solutions across diverse industries.