Tokyo, November 28, 2022 - NEC Corporation (NEC; TSE: 6701) has developed a highly accurate technology that recognizes in real time a wide variety of work activities performed by multiple people from images captured by cameras. Even in congested environments with many people, such as construction sites, which have been difficult to accurately analyze until now, it is possible to stably recognize various work activities at one time. Daiwa House Industry Co., Ltd. (Daiwa House Industry) and NEC recently confirmed the effectiveness of this technology in joint demonstration tests at a construction site.

This technology analyzes the relationships between various features of humans and objects with different representations. In addition to the visual features of humans and objects, it also analyzes the relationships between features of human poses, class information for objects such as tools and heavy machinery, the position of humans or objects, as well as visual features of the surrounding environment. Moreover, this technology adaptively weighs important features by deep learning as it recognizes each work activity. This makes it possible to accurately recognize a wide variety of work activities, even in congested environments where many people interact, such as construction sites.

In the future, in addition to construction, NEC will proceed with verification of this technology for various on-site operations, including manufacturing, logistics, and retail sites, while aiming for commercialization in fiscal 2023.

Main features of the technology include the following:
1. Simultaneous recognition of a wide variety of work activities
At construction sites, many people simultaneously perform a wide variety of work activities using various tools and heavy machinery with various actions and postures to interact with the surrounding environment. In addition to the visual features of humans and objects, this technology integrates and analyzes the relationships between various features with different representations, such as human poses, object class information, the positions of humans and objects, and visual features of surrounding environments, which is important for recognizing the content of an activity. This makes it possible to simultaneously recognize a wide variety of work activities, such as “excavation," “compaction," “leveling," “concrete leveling," “concrete pouring," “cart transport," and “rebar assembly" performed at construction sites.

2. Stable recognition of work activities, even in crowded environments
Conventionally, it was difficult to accurately recognize the activities of workers in crowded environments, such as construction sites. This technology adaptively weighs important features by deep learning when analyzing relationships between various features, so that even if some features are not detected due to human overlap, etc., they can be complemented by other features. Even when a large number of people are crowded together, it is possible to accurately identify the relationship between a human and an object by focusing on position information. This makes it possible to stably recognize work activity in a crowded environment.

Outline of the joint verification tests
NEC, in collaboration with Daiwa House Industry Co., Ltd., conducted demonstration tests of this technology from March to May 2022 at a construction site of detached houses. As a result, it was possible to accurately recognize multiple work activities, such as "compaction," "root cutting and backfilling," "concrete placement," and "rebar assembly," and to estimate the work time of each activity by field workers with less than a 10% error.

***

About NEC Corporation
NEC Corporation has established itself as a leader in the integration of IT and network technologies while promoting the brand statement of “Orchestrating a brighter world.” NEC enables businesses and communities to adapt to rapid changes taking place in both society and the market as it provides for the social values of safety, security, fairness and efficiency to promote a more sustainable world where everyone has the chance to reach their full potential. For more information, visit NEC at https://www.nec.com.