| Summary: | 碩士 === 國立中正大學 === 電機工程研究所 === 107 === Architecture is one of the most important manifestations of human cultural
characteristics. We can understand the customs and spirit of each culture from the
architectural style and understand its aesthetic. However, with the evolution of time,
war, natural disasters and various man-made destruction. Many characteristic historic
buildings have been damaged to a considerable extent. How to reconstruct and repair
has become a problem for archaeological and architectural fields to solve for many
years.
Artificial Intelligence is a technology that has been rapidly increasing in
discussion over the past few years. With the improvement of hardware, many
applications have been developed on mobile devices. In the field of image processing,
AI is often used in image enhancement to make photos more beautiful, improving the
proportion or color of individual images for the public to accept. In this paper, I use
machine learning to inpaint inpaired images of building. The Generative Adversarial
Network with various improvements can produce more accurate inpainted images on
the quantified data, making the inpainting follow an effective reference.
|
|---|
