项目作者: keswani-Rohitkumar

项目描述 :
Image Super-resolution Using Deep Learning
高级语言: Python
项目地址: git://github.com/keswani-Rohitkumar/Image_Super-Resolution_using_CNN.git


Image_Super-Resolution_using_CNN

Image Super-resolution Using Deep Learning

Deep Convolutional Model is superior to perform image super-resolution because SRCNN achieves the highest PSNR (Peak Signal to Noise Ratio). It is a ratio of the maximum possible power of a signal and the power of corrupting noise that affects the fidelity of its representation.The deep convolutional model in SRCNN directly learns anend-to-end mapping between low and high-resolution images with little preprocessing. This model has achieved superior performance than the state of art methods. It is also believed that more performance can be achieved by experimenting withmore filters and different strategies. Moreover, with the robustness and simplicity of the model it canalsobe used in various low-level vision problems.

PSNR is the Peak signal-to-noise ratio (PSNR) is defined as theratio ofthe maximum possible power of a signal and the power of corrupting noise that affects the fidelity of its representation.If the value of the PSNR is high the betteristhe model to reconstruct a high-resolution image from a low resolution image.

Ground-Truth Image

image

HR-BI (PSNR =20.497630181368823)

image

HR-SRCNN (PSNR=22.922696428588342)

image

(PSNR for HR_image and LR_image is : 20.497630181368823

PSNR for HR_image and SR_image is : 22.922696428588342

)