Newsroom

<figure class="image"><img style="aspect-ratio:640/358;" src="https://deeply-website-backend.s3.ap-northeast-2.amazonaws.com/deeply/2024-11/817a1b99-80fa-478c-a482-f9b964e0a346.png" width="640" height="358"></figure>&nbsp;On the 20th, sound analysis AI company, "DeeplyAI", founded by the CEO, Suji Lee, announced the installation of ten units of their emergency detection solution, “ListenAI,” in the restrooms of Incheon National University Station. &nbsp;This initiative is part of a project led by Incheon Transit Corporation called the 'Demonstration of Emergency Situation Detection for Subway Restrooms Using Sound-Based AI.' With the increasing proportion of elderly users in urban transit, concerns about the possibility of emergency situations due to health issues have been continuously raised, alongside growing public fears of violent crimes occurring in restrooms. &nbsp;Incheon Transit Corporation designed this project to establish a system capable of rapid situation assessment and response through the use of AI. ListenAI has now become a core resource in restroom safety and security at Incheon National University Station. Previously, restroom safety relied primarily on public reporting and routine staff checks. With the implementation of ListenAI, however, real-time monitoring has become possible. ListenAI detects sounds on-site in real-time, and if it identifies an abnormal situation, it immediately triggers an alarm. It provides the station office with analysis results including the time, location, and type of incident, allowing station staff to respond swiftly. Suji Lee, CEO of Deeply, explained, “ListenAI combines high analytical accuracy with a lightweight model that can run smoothly on low-power edge servers, like those in smartphones. This makes it accessible to organizations without high-performance servers. ListenAI can also distinguish critical sounds in noisy environments, making it a 24/7 emergency monitoring solution for diverse settings, such as construction sites, factories, and public streets.”&nbsp;<a href="https://zdnet.co.kr/view/?no=20240920114643">https://zdnet.co.kr/view/?no=20240920114643</a>

Deeply and Incheon Transit Corporation Install ‘ListenAI’ in Incheon National University Station Restrooms

관리자

[Digital Future Innovation Awards: KCA Director's Award] Deeply’s 'ListenAI'&nbsp;“We want to be a company that contributes meaningfully to the world by creating better technology and products”, Suji Lee, CEO of Deeply, a sound analysis AI startup, made a remark after receiving the KCA President's Award at the 8th Korea Digital Future Innovation Awards, hosted by the Ministry of Science and ICT and organized by the Korea Software Industry Association (KOSA).&nbsp;Founded in 2017 with the mission of "adding meaning to sound," Deeply has built a reputation as a technology-driven startup specializing in sound analysis. Starting with a service that could analyze baby cries to determine their cause, Deeply has evolved to focus on detecting abnormal sounds from machinery in manufacturing facilities and analyzing emergency-related sounds for enhanced safety.&nbsp;<figure class="image"><img style="aspect-ratio:1918/1019;" src="https://deeply-website-backend.s3.ap-northeast-2.amazonaws.com/deeply/2024-11/bcbda726-da4f-4cae-b9e5-2f5e5693da88.png" width="1918" height="1019"></figure>ListenAI Installed at the Government Sejong Complex Sports Center (Photo by Deeply)&nbsp;The award-winning production, "ListenAI," goes beyond the traditional CCTV monitoring by analyzing sounds within the scene to immediately alert control rooms of potentially dangerous incidents. While multiple CCTV cameras record footage across facilities, they are generally used only in post-incident reviews. ListenAI, however, provides real-time detection of collision sounds, explosions, screams (distinguished by gender), and noises indicating fights or disturbances, directly notifying facility managers of ongoing threats.&nbsp;This AI solution helps prevent incidents that might otherwise go unnoticed due to limited personnel. It also allows for quick coordination with police and medical services, covering areas not visible to cameras, low-light situations, and locations difficult to monitor with video alone.&nbsp;Sound analysis technology, which includes real-time analysis of audio sources, is in its early stages domestically and is known to be complex due to factors like noise and location variability. Deeply has overcome these challenges to commercialize a solution robust enough for field use, making it possible to analyze sound patterns and detect incidents in environments with mixed and overlapping sounds.&nbsp;This year, Deeply’s contributions have also been recognized at ICASSP 2024 with a published paper showcasing their advanced sound recognition technology, which analyzes sounds from multiple microphones to assess situations at one of the highest levels globally. Additionally, they hold a versatile deep-learning patent titled “Real-Time Sound Analysis Method and System.”&nbsp;— Heejung Ahn, ZDNet Korea<a href="https://zdnet.co.kr/view/?no=20241007130920">https://zdnet.co.kr/view/?no=20241007130920</a>&nbsp;

Deeply: "We Aim to Make a Positive Impact on the World by Analyzing Sound"

메인관리자

<figure class="image"><img style="aspect-ratio:950/531;" src="https://deeply-website-backend.s3.ap-northeast-2.amazonaws.com/deeply/2024-11/c4768331-1ee6-4605-be5f-580184e7c30e.png" width="950" height="531"></figure>&nbsp;Deeply, a company specializing in sound analysis artificial intelligence engines, announced on the 23rd that it has published a research paper at ICASSP 2024, an international signal processing conference.ICASSP, organized annually by the Institute of Electrical and Electronics Engineers (IEEE), is the world's largest international conference in the fields of speech, image, communication, and signal processing. This year marks the 49th conference and, for the first time, it was held in South Korea. The event attracted global companies like Apple and Google, along with over 4,000 researchers.&nbsp;Deeply’s paper demonstrates advancements in sound analysis AI, particularly addressing the issue of decreased accuracy due to frequency characteristic differences across various recording devices. The solution uses a proprietary acoustic transformation technology to generate data for AI training, which has led to improvements in generalization performance (the AI's ability to analyze new data) from 5.2% to 11.5%.&nbsp;The paper highlights an acoustic transformation model built on the 'CycleGAN', a deep learning algorithm for modifying the attributes of media such as images, sound, and text. The model was further refined with Deeply’s unique data augmentation techniques.To address variability in recording equipment, Deeply trained the model on the acoustic properties of various devices, applying different frequency patterns to their existing sound data. This approach allowed them to simulate the effect of recording the same sound on multiple devices, enriching the data with diverse frequency patterns.&nbsp;Deeply's sound analysis AI engine, after refinement using the generated data, maintains high accuracy even in new environments. The engine has been trained on over 50,000 hours of sound data to reach high levels of precision.Already recognized for its ability to detect subtle sounds, the technology is gaining attention in security sectors where intruder detection is critical. It is also applied in the manufacturing industry for detecting equipment malfunctions and product defects.&nbsp;Suji Lee, CEO of Deeply, stated, “Our AI sound analysis engine’s accuracy ranks among the highest globally. We aim to continue expanding our capabilities, establishing Deeply as a leader in sound analysis solutions that serve diverse needs across society.”&nbsp;[Unicorn Factory, MoneyToday Startup Media Platform]<a href="https://news.mt.co.kr/mtview.php?no=2024072317041758293">https://news.mt.co.kr/mtview.php?no=2024072317041758293</a>