A Dynamic Study of Art Image Digitization Based on Metadata

With the continuous development of the internet age, more and more art images are taking on digital forms, resulting in a new way of survival for art image digitization. However, the digitization process of art images is affected by various factors, resulting in poor results and low digital quality of art images. Therefore, this article conducted research on the digitization of art images based on metadata, and utilized BP (Back Propagation) neural network for metadata processing and analysis to achieve metadata visualization and interactive design. Animation production software was then utilized for image compression, transparent display, and modeling, and finally interactive display technology was used to display the dynamic design of art images. 4000 user feedback data and art image metadata from four age groups were collected and named A art image set. Starting from the visual communication effect, accuracy, and fidelity of art images, the differences in dynamic design of A art image digitization were compared. The experimental results showed that 2820 people were satisfied with the visual communication effect of dynamically designed art images, with a satisfaction rate of 70.5%. Only 1070 people in the control group were satisfied. The metadata accuracy of dynamically designed art images was greater than 80%, and the average accuracy was close to the median line, with small overall fluctuations. The deviation value between dynamically designed art image data and standard images is small, and the overall fidelity is relatively high. In short, the evaluation effect of digital dynamic design of art images is very good.