#docomonejp
**NTT Docomo Develops Privacy-Protecting Technology to Visualize Behavior and Interests from Video Footage**
NTT Docomo has developed a technology that visualizes people’s behavior and interests from videos taken by smartphones and surveillance cameras, while removing personal information. The technology extracts necessary data such as skeletal information from the videos, then removes personal information like faces and clothing, and draws the person’s skeleton with lines. This allows visualization of people’s behavior while protecting privacy.
The technology also automatically generates the background of the removed person, resulting in more natural images compared to processed images like mosaics. The original videos are sent to a high-security “docomo MEC” network and processed within the service, reducing the risk of information leakage during cyber attacks while analyzing people’s behavior.
The technology can project information such as stay time, gaze, and object touch onto a 3D virtual space created from the video. For example, when used in convenience stores or commercial facilities, it can count customers, estimate gender and age, and obtain heatmaps of customers’ product touches, gazes, and in-store movement. This data can help identify easily picked-up products, high-traffic areas, and products often compared at purchase, allowing for improvements in store marketing.
By combining the technology with the metaverse, it can be applied to various services such as theme parks where people’s real-world behavior is transferred to a virtual space. Starting in April, Kokuyo and NTT Communications will conduct a demonstration experiment on visualizing people’s behavior in office spaces using this technology. They will analyze employees’ movement and behavior within Kokuyo’s office and verify the effectiveness of the technology. NTT Docomo aims to commercialize the technology by the end of the fiscal year 2023.
https://bizzine.jp/article/detail/9135
https://bizzine.jp/article/detail/9135