Design and Application of Hybrid Collaborative Interaction Model for Game Scenarios

ABSTRACT

With the swift advancement of the gaming industry, players' demands for interactive gaming experiences are increasing day by day. Traditional single interaction modes are no longer able to meet the complex and diverse demands of gaming scenarios. In this context, the research designed a hybrid collaborative interaction model tailored for game scenes. A human‐computer interaction gesture recognition network was constructed based on the Masked Region Convolutional Neural Network. Feature extraction and region segmentation operations were carried out using a Feature Pyramid Network and a Region Proposal Network, among others. Subsequently, a four‐layer hybrid collaborative interaction model was designed, integrating devices such as HoloLens, to enable multi‐modal interaction encompassing gestures, voice, and visuals. The experiment outcomes indicated that the gesture recognition network had an accuracy rate of 98.3%, a loss rate of 3.6%, an average accuracy of 50.3%, and a processing speed of 78.5 frame/s. The average running stability of the game scene hybrid collaborative interaction model was 98.8%, the response stability was 99.3%, the accuracy of gesture, speech, and visual recognition exceeded 97%, and the scene synchronization accuracy was 99.1%. The performance of the designed model was excellent, with precise understanding of user intentions and good collaboration. The research provides feasible solutions for the innovation of game interaction modes, helping the gaming industry develop towards more immersive and diversified directions.