BiCrossNet with Decoupled Dual Generators: A Parameter‑Efficient and Generalizable Few‑Shot Custom Gesture Recognition Framework

Gesture interaction and touchless sensing is a natural and intuitive control method that allows users to control devices through natural hand or body movements, reducing reliance on physical input and enhancing convenience and functional efficiency. The biggest challenge is balancing accuracy and model parameter count for custom gesture tasks, while minimizing the number of data samples required for training. This paper proposes a novel IMU-based BiCrossNet model and two novel data augmentation models (delta-generator and embedding-generator) to address this challenge. Compared with existing methods, the model proposed in this article boosts accuracy by 11.7% and 12.7% in UMAHand (public datasets); and 8.85% and 5.25% in GRHand (self-developed datasets), with decreasing 27.8% pretrained feature-extractor model parameters. This research lays a solid foundation for deploying and implementing custome gesture recognition engineering on intelligent terminal devices.

Gesture interaction and touchless sensing is a natural and intuitive control method that allows users to operate devices through natural hand or body movements, reducing reliance on physical input and enhancing convenience and functional efficiency. However, existing gesture recognition technologies face a critical challenge: they are either not accurate enough or too large to run on edge devices, while requiring huge amounts of training data to work well. To solve this problem, we developed a new gesture recognition model and two innovative data generation techniques. Our experiments show that our model significantly improves accuracy on both public and our own datasets, while considerably reducing model size. This research will make it much easier to deploy gesture recognition technology on everyday edge devices, bringing more convenient and intuitive interaction experiences to users worldwide.