NEC Corporation (NEC) announced on the 29th that it has developed a new technology. It uses video analysis and generative AI to find gaps between “model” movements and actual performance. It also gives automated improvement recommendations.
This innovation consists of two core technologies: (1) a video analysis system that detects subtle deviations from the model movement, and (2) a generative AI system that translates these differences into precise advisory feedback. By replacing traditional instructor-led guidance with AI-generated advice, NEC aims to facilitate self-learning across various job functions.
The technology compares a model’s movements to an employee’s work by aligning segments. Unlike conventional methods, this system tracks both movements and interactions with objects, like “grabbing” and “holding.” It ensures precise matching, even if timing varies. This capability allows for the detection of nuanced differences that were previously undetectable.
Also Read: KIOXIA AiSAQ™ Tech to Cut DRAM in AI Systems Released Open Source
The generative AI component enhances this analysis by processing video data along with skeletal information, including hip and knee movements and hand and finger positioning. By integrating this comprehensive data into a large-scale vision and language model (VLM), the AI accurately assesses posture and movement, generating specific and actionable improvement recommendations.
Furthermore, by presenting both the advisory feedback and corresponding video clips, employees can refine their skills independently, reducing reliance on instructors. This technology is particularly beneficial in industries requiring precision, such as detailed assembly work, packaging, and logistics. Additionally, it contributes to lowering training costs.
NEC plans to conduct demonstrations and advance product development throughout fiscal 2025, with the goal of launching the service by fiscal 2026.
