MMES
Manifold Modeling in Embedded Space (MMES): An Interpretable Alternative to Deep Image Prior
[IEEE Transactions on Neural Networks and Learning Systems]
Please download our paper from IEEE (Open Access) : https://ieeexplore.ieee.org/abstract/document/9281370
Link to arXiv paper : https://arxiv.org/abs/1908.02995
It is implemented by using TensorFlow (please try our codes for image inpainting, super-resolution, deconvolution, and denoising)
-
demo_MMES_IP(inpainting) recovers the following incomplete image with missing entries (70% pixels are missing)
-
demo_MMES_SR(super-resolution) recovers the following low resolution image (128 x 128) to (512 x 512)
-
demo_MMES_DC(deconvolution) recovers the following blurred image with Gaussian kernel
-
demo_MMES_DN(denoising) recovers the following noisy image
-
Additional Inpainting Demo. for gray-scale image and binary image were uploaded