This study incorporates the smart learning features with intelligent image processing technology in developing an artificial intelligence stylized painting system. By employing a color transformation model and a smart learning algorithm, the system is able to provide intelligent color matching and continuous stylistic progression. In color matching experiments, the system has achieved PSNR scores between 31.22 dB and 33.48 dB, while its SSIM values are 0.9852 and 0.9925. The results show a clear superiority compared to various conventional classification algorithms. In terms of the quality of generated images, the system’s PSNR scores for pictures produced under various themes range between 75.83 dB and 78.72 dB, while their SSIM scores exceed 0.97. This shows significant superiority over other comparison models utilizing style translation and texture synthesis methods. In addition, the system has shown excellent performance in line processing by consuming less than 4.97 ms in processing highly intricate lines. In terms of resource utilization, the system still uses only 45.45% GPU resources and 19.5% CPU usage for generating 500 images – much lower than those of the comparison approaches.
