Mr. Sangjoon Han, CTOThis is where NSENSE introduces its breakthrough solution.
NSENSE’s ultra-light AI model, developed after years of extensive research, can operate even on edge devices such as mini PCs without losing high accuracy. To achieve this, the company went beyond just algorithmic optimization of AI inference engines to improve the AI model by applying the latest AI technologies, such as quantification, model pruning, and knowledge distillation transfer learning. NSENSE’s ultra-light AI cognitive model can operate using CPU or NPU instead of using the commonly used gaming GPU to infer the AI cognitive model.
NSENSE is driven by the mission to develop technology that makes life easier by researching the N-th senses that go beyond the human senses. To this end, the company has introduced AR technology connecting virtual reality through ‘ARme’ and ‘NVIS’ solutions. In addition, NSENSE’s ‘PTAS’ solution can replace human visual intelligence, especially in the market of CCTV solutions for security surveillance and customer analysis.
The Next Revolution in Facial Recognition Technology
NSENSE offers its AI cognitive technology to clients who need CCTV video analysis. These solutions can be used to identify customer behavior in supermarkets, analyze data, or protect residents’ safety in residential complexes. Generally, these environments require users to explore dozens of CCTVs in real-time, and simultaneously, the high- precision AI cognitive model requires vast computing power. While many companies tried to overcome this by applying multi-GPU technology or using expensive GPUs, they naturally lead to high initial costs.
“Unlike other solutions that use multi-GPU technology or expensive GPUs, NSENSE’s AI cognitive model analyzes dozens of CCTVs in real-time in a cost-efficient manner, without negatively impacting the accuracy,” says Mr. Sangjoon Han, the CTO of NSENSE.
The solution can recognize a customer’s face and even analyze attributes such as gender, age, and facial expression to provide detailed, statistical information. For instance, several large retail stores today leverage NSENSE’s solution to statistically analyze the number of customer visits at different time zones. This plays a vital role in improving the operational efficiency of a store. In addition, by recognizing the visitors’ faces and generating a vectorized digital fingerprint, store operators can calculate the frequency of visits for each visitor to see if they are new or existing customers. This allows them to better understand the effectiveness of a promotion, evaluate customer loyalty through the frequency of revisiting, and gain insights into setting long-term goals for stores. Also, the solution informs store staff through push messages in case visitors are wandering around the store or looking for products. In such a case, store employees can quickly access visitors who need guidance to induce shopping— ultimately increasing customer satisfaction and store profits.
Apart from the retail industry, several housing management companies utilize NSENSE’s solution to separate residents from outside visitors through facial recognition. If someone loiters or attempts to enter an area where entry is prohibited, the solution can notify the administrators and prompt them to take appropriate action.
In a nutshell, NSENSE’s facial recognition technology shows excellent performance in analyzing face information and 1:1 face verification and 1:N face identification with high accuracy.
"Our Clients Are Indirectly Benefiting From Our Work Culture. It Reduces Risks And Brings Tangible Results"
Data Security - Never an Afterthought
Other than its powerful solution, what makes NSENSE stand out in the market is the company’s continuous focus on data security. NSENSE understands that an unspecified number of people can be filmed on CCTV, and a person’s face is a piece of unchanging biometric information. So, data breaches involving facial recognition data can increase the potential for identity theft, stalking, and harassment.
NSENSE holds itself responsible for protecting such personal information from being leaked. It only sends analyzed information, such as statistical information, to the cloud server and not the CCTV videos. Image analysis is accessed only from edge devices on a local network, and edge devices are equipped with AI cognitive models to perform analysis independently without external connections. This reduces the risk of leakage of CCTV images or the personal information of a person.
The solution can recognize a customer’s face and even analyze attributes such as gender, age, and facial expression to provide detailed, statistical information. For instance, several large retail stores today leverage NSENSE’s solution to statistically analyze the number of customer visits at different time zones. This plays a vital role in improving the operational efficiency of a store. In addition, by recognizing the visitors’ faces and generating a vectorized digital fingerprint, store operators can calculate the frequency of visits for each visitor to see if they are new or existing customers. This allows them to better understand the effectiveness of a promotion, evaluate customer loyalty through the frequency of revisiting, and gain insights into setting long-term goals for stores. Also, the solution informs store staff through push messages in case visitors are wandering around the store or looking for products. In such a case, store employees can quickly access visitors who need guidance to induce shopping— ultimately increasing customer satisfaction and store profits.
Apart from the retail industry, several housing management companies utilize NSENSE’s solution to separate residents from outside visitors through facial recognition. If someone loiters or attempts to enter an area where entry is prohibited, the solution can notify the administrators and prompt them to take appropriate action.
In a nutshell, NSENSE’s facial recognition technology shows excellent performance in analyzing face information and 1:1 face verification and 1:N face identification with high accuracy.
"Our Clients Are Indirectly Benefiting From Our Work Culture. It Reduces Risks And Brings Tangible Results"
Data Security - Never an Afterthought
Other than its powerful solution, what makes NSENSE stand out in the market is the company’s continuous focus on data security. NSENSE understands that an unspecified number of people can be filmed on CCTV, and a person’s face is a piece of unchanging biometric information. So, data breaches involving facial recognition data can increase the potential for identity theft, stalking, and harassment.
NSENSE holds itself responsible for protecting such personal information from being leaked. It only sends analyzed information, such as statistical information, to the cloud server and not the CCTV videos. Image analysis is accessed only from edge devices on a local network, and edge devices are equipped with AI cognitive models to perform analysis independently without external connections. This reduces the risk of leakage of CCTV images or the personal information of a person.
The heightened focus on security can make NSENSE’s solution highly impactful in the entire APAC region, where many countries are introducing stricter privacy protection regulations and restricting the storage and handling of personal information.
A work Culture that Complements the Vision
NSENSE uses the Scrum development methodology. All projects write epics and backlogs with the participation of developers and estimate the speed of development with story points. The company conducts development through Sprint every two weeks and decides whether to distribute the results through peer evaluation and reflect on the next Sprint. This enables quicker product release while maintaining higher quality and improving customer satisfaction.
A prominent driving force behind NSENSE’s success is its continued efforts to develop technologies and engage employees such that they trust and follow the company’s larger vision. It has created an environment that improves the work culture and propels research and development. NSENSE enables its researchers to decide their working hours, allowing them to work beyond the limits of the place.
“Our clients are indirectly benefiting from this work culture. It reduces risks and brings tangible results,” says Mr. Han. The company’s research institute was evaluated as a work innovation grade A in Korea in 2021 and was adjudged as a workplace innovation excellent company.
Geared for the Future
NSENSE plans to raise the accuracy of face recognition technology to the world’s highest level in the next few years and, consequently, develop a solution that can protect the lives and safety of workers in factories and industrial sites. Not stopping there, the company is researching a payment method that can authenticate facial identity even if a visitor is wearing a mask in unstaffed stores.
“We may be a small startup from an external perspective, but we are proud of our AI cognitive technology,” states Mr. Han. “We have signed sales agency contracts with several large Japanese companies. They listen to and analyze clients’ needs and provide insights to build our solution portfolio correctly. Planning products through actual clients’ voices are helpful to us.”
NSENSE has recently raised two rounds of investment from venture capital firms in Singapore and is preparing to be listed on the Japanese stock market in two years. Expanding sales channels to the APAC region, including Korea, is also in NSENSE’s pipeline.
Being an innovator in the field, NSENSE is confident about the power of AI cognitive technology in driving the world toward a promising and safe future. Undoubtedly, NSENSE will play a big part in that movement, especially in the APAC region.
A work Culture that Complements the Vision
NSENSE uses the Scrum development methodology. All projects write epics and backlogs with the participation of developers and estimate the speed of development with story points. The company conducts development through Sprint every two weeks and decides whether to distribute the results through peer evaluation and reflect on the next Sprint. This enables quicker product release while maintaining higher quality and improving customer satisfaction.
A prominent driving force behind NSENSE’s success is its continued efforts to develop technologies and engage employees such that they trust and follow the company’s larger vision. It has created an environment that improves the work culture and propels research and development. NSENSE enables its researchers to decide their working hours, allowing them to work beyond the limits of the place.
“Our clients are indirectly benefiting from this work culture. It reduces risks and brings tangible results,” says Mr. Han. The company’s research institute was evaluated as a work innovation grade A in Korea in 2021 and was adjudged as a workplace innovation excellent company.
Geared for the Future
NSENSE plans to raise the accuracy of face recognition technology to the world’s highest level in the next few years and, consequently, develop a solution that can protect the lives and safety of workers in factories and industrial sites. Not stopping there, the company is researching a payment method that can authenticate facial identity even if a visitor is wearing a mask in unstaffed stores.
“We may be a small startup from an external perspective, but we are proud of our AI cognitive technology,” states Mr. Han. “We have signed sales agency contracts with several large Japanese companies. They listen to and analyze clients’ needs and provide insights to build our solution portfolio correctly. Planning products through actual clients’ voices are helpful to us.”
NSENSE has recently raised two rounds of investment from venture capital firms in Singapore and is preparing to be listed on the Japanese stock market in two years. Expanding sales channels to the APAC region, including Korea, is also in NSENSE’s pipeline.
Unlike Other Solutions That Use Multi-Gpu Technology Or Expensive Gpus, Nsense’s Ai Cognitive Model Analyzes Dozens Of Cctvs In Real-Time In A Cost-Efficient Manner, Without Negatively Impacting The Accuracy
