Blog: Loss Function For Super Image Super Resolution
In Super-Image-Super-Resolution (SISR), Autoencoder and U-Net are heavily used; however, they are well-known for difficulties in training to convergence. The choice of loss function plays an important role in guiding models to optimum. Today, I introduce 2 loss functions for Single-Image-Super-Resolution that is based on Convolution and Sobel operator:
- Mean Squared Error (MSE)
- Mean Gradient Error (MGE)
– Please follow this link to see the full blog of the topic –
References
[1] Single Image Super Resolution based on a Modified U-net with Mixed Gradient Loss, https://arxiv.org/pdf/1911.09428.pdf
[2] Design of an image edge detection filter using the Sobel operator, https://ieeexplore.ieee.org/document/996