Artistic Image Style Transfer Based on Laplacian Pyramid Network

As a very creative research direction in the field of deep learning, style transfer has been attracted lot attention past two years, which can transform ordinary photos into artistic paintings and generate interesting pictures. Existing methods be divided categories: arbitrary single transfer. The representative method LapStyle synthesizes stylized image progressive procedure by using AdaIN, an method. Inspired it, this paper proposes novel framework named LapAttN to optimize transfer, combines advantages AdaAttN. experimental results on COCO dataset show that quality proposed is superior other SOTA algorithms.