Research on the Effect Evaluation of Function Vision System Algorithm in Computer Experiments under Virtual Simulation Environment

This study describes an innovative visual recognition system based on the regional convolutional neural network (R-CNN) architecture, with the core goal of optimizing the accuracy and efficiency of target detection and identification. In the initial stage, a test platform customized for virtual simulation scenarios was built to simulate and verify the algorithm. This algorithm uses convolutional neural networks to mine image features and estimates target regions through region proposal networks (RPNs) to achieve accurate positioning and classification. To confirm the effectiveness of the algorithm, this study conducted a series of experiments covering tests of various virtual scenes and target types. Experimental data revealed that the innovative algorithm showed outstanding advantages over traditional methods in target recognition accuracy and classification rate, with an average accuracy of 92.7% and a processing rate gain of 15.3%. Furthermore, through detailed data analysis, this study evaluated the performance of the algorithm under different parameter configurations and selected the optimal model settings, thereby enhancing the stability and adaptability of the algorithm. In summary, the application of this algorithm in a virtual simulation environment shows high practical value and broad application prospects.