A Literature Review of Image Stylization Techniques
Abstract
Image stylization is process to represent content of image in different styles. Non-Photorealistic Rendering (NPR) is an area where various techniques are evolved to transfer image into different artistic styles. In recent years, power of convolution neural network (CNN) shows its excellence in image stylization techniques which is known as Neural Style Transfer (NST). Subsequently it receives lots of attention of researcher in the field of NST to extend and improve different approaches. This review paper, summarizes both NPR and NST techniques for image stylization. Firstly, we studied the present techniques in the area NPR of NST. Then, it describes photorealistic stylization techniques and evaluation methods to compare image stylization techniques. Further study extends with a different challenges of image stylization techniques and future research directive.