The digital era has brought unprecedented convenience and prosperity to art design. This project develops the application of artificial intelligence technology from the digitalization process of traditional visual art elements in two dimensions: image enhancement and image recognition. Combining the quantile algorithm and adaptive gamma correction technology, we have designed an image enhancement method based on AGC-Quantile for traditional elements. Based on the classification model of the MobileNet network and the target detection model based on YOLOv5, the traditional element image recognition model is proposed. Traditional element images containing multiple types of tattoos are selected as experimental objects, and the actual performance of this model is investigated through algorithm comparison and evaluation index analysis. The AGC-Quantile method processes different tattoo images with high quality and outperforms the comparison algorithms in different indexes (MSE>17, PSNR>23, and SSIM>0.82). The classification model based on the MobileNet network and the target detection model based on YOLOv5 is proposed to enhance traditional element images. -YOLOv5 has a classification accuracy and target detection accuracy of 83.2% and 87.6%, respectively, and improves 65.82% in time. According to the experiment, the AI-based model constructed in this paper has superior image enhancement and image recognition effects and can be utilized for the digitization of traditional visual art elements.