The DIV2K dataset is used for training, which has a high quality 2K resolution dataset for image restoration tasks. Result of this architecture from the research paper- Source: Code implementation of ESRGAN Replaces original basic blocks with a proposed residual-in-residual dense block, which combines multi-level residual network and dense connection as shown in the above figure. ![]() The above fig shows the proposed architectures in this approach, the authors try to improve the quality of the recovered image given SRGAN by doing two main modifications in the generator’s structure: 1. ![]() The generator uses a linear combination of Perceptual difference between real and fake images using a pre-trained VGG19 network, Pixel wise absolute difference between real and fake images, and Relativistic average loss between real and fake images function during adversarial training. The ESRGAN uses a Relativistic discriminator to better approximate the probability of an image being real or fake thus, the intern produces better results. This allows the ESRGAN to produce images with a higher approximation of the sharp edges of the image artifacts. In addition to that, the model lacks a batch normalization layer in the generator to prevent smoothing out the artifacts in the images. The model uses Residual-in-Residual block as a basic convolution block instead of a basic residual network or simple convolution trunk to provide a better flow gradient at the microscopic level. Today in this article, we will discuss the Enhanced Super Resolution GAN, an improved version of Super-Resolution GAN and its python code implementation. These PSNR oriented approaches tend to make output over-smoothed without sufficient high-frequency details the PSNR metric fundamentally disagrees with the subjective evaluation of human observers. When the two images are identical, the absence of noise says the mean squared error is zero for this case, PSNR is infinite. ![]() Thus, interpretation can be made like, higher the PSNR value better the reconstruction. Typical values of PSNR in lossy images vary between 30db to 50db, provided the bit depth is 8 bits. Thus, while comparing the compressed images, PSNR is an approximation to human perception of reconstruction quality. The signal, in this case, is original image data noise is the error introduced by compression. PSNR is the most commonly used measure that measures the quality of reconstruction of lossy compressed images.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |