Deblurring from a motion blurred image has been studied for some times. Recently, convolution neural network(CNN) has been used widely and it can be used on finding the blur kernel or the latent sharp edge of a blurred image. In recent years, the generative adversarial network (GAN) performs well on style transformation. We consider that a deblurring problem as a style transformation problem. We focus on improving the DeblurGAN\'s generator, which is the state-of-the-art of the deblurring method and present a new kind of block which combined inception block, residual block and dense block to do deblurring from motion blur. By using the conception of DenseNet which can avoid overfitting. The improved DeblurGAN presents better in both structural similarity measure and by visual effect
• Hardware: Processor: i3 ,i5 RAM: 4GB Hard disk: 16 GB • Software: operating System : Windws2000/XP/7/8/10 Anaconda,jupyter,spyder,flask Frontend :-python Backend:- MYSQL
₹10000 (INR)
2019