Recursive quantitative analysis modelling of computer art design interaction

With the increasing demand for high-quality images in computer art interaction design models, image compression is crucial for effective transmission of information. This study combines traditional recursive characteristics with convolutional neural network models to propose a recursive convolutional neural network-based image compression algorithm. The algorithm's performance was tested on Kodak1 and Kodak2 datasets, showing a decreasing trend in mean square error as the number of iterations increases. At 800 iterations, the algorithm outperformed other algorithms in terms of mean square error. Additionally, it exhibited higher peak signal-to-noise ratio and multi-scale structural similarity compared to traditional neural network algorithms. These results indicate that the proposed algorithm effectively compresses images and efficiently handles the large volume of images generated in artistic interaction design.