A. Fakhari; K. Kiani
Abstract
Image restoration and its different variations are important topics in low-level image processing. One of the main challenges in image restoration is dependency of current methods to the corruption characteristics. In this paper, we have proposed an image restoration architecture that enables us to address ...
Read More
Image restoration and its different variations are important topics in low-level image processing. One of the main challenges in image restoration is dependency of current methods to the corruption characteristics. In this paper, we have proposed an image restoration architecture that enables us to address different types of corruption, regardless of type, amount and location. The main intuition behind our approach is restoring original images from abstracted perceptual features. Using an encoder-decoder architecture, image restoration can be defined as an image transformation task. Abstraction of perceptual features is done in the encoder part of the model and determines the sampling point within original images' Probability Density Function (PDF). The PDF of original images is learned in the decoder section by using a Generative Adversarial Network (GAN) that receives the sampling point from the encoder part. Concretely, sampling from the learned PDF restores original image from its corrupted version. Pretrained network extracts perceptual features and Restricted Boltzmann Machine (RBM) makes the abstraction over them in the encoder section. By developing a new algorithm for training the RBM, the features of the corrupted images have been refined. In the decoder, the Generator network restores original images from abstracted perceptual features while Discriminator determines how good the restoration result is. The proposed approach has been compared with both traditional approaches like BM3D and with modern deep models like IRCNN and NCSR. We have also considered three different categories of corruption including denoising, inpainting and deblurring. Experimental results confirm performance of the model.
H.5.11. Image Representation
E. Sahragard; H. Farsi; S. Mohammadzadeh
Abstract
The aim of image restoration is to obtain a higher quality desired image from a degraded image. In this strategy, an image inpainting method fills the degraded or lost area of the image by appropriate information. This is performed in such a way so that the obtained image is undistinguishable for a casual ...
Read More
The aim of image restoration is to obtain a higher quality desired image from a degraded image. In this strategy, an image inpainting method fills the degraded or lost area of the image by appropriate information. This is performed in such a way so that the obtained image is undistinguishable for a casual person who is unfamiliar with the original image. In this paper, different images are degraded by two procedures; one is to blur and to add noise to the original image, and the other one is to lose a percentage of the pixels belonging to the original image. Then, the degraded image is restored by the proposed method and also two state-of-art methods. For image restoration, it is required to use optimization methods. In this paper, we use a linear restoration method based on the total variation regularizer. The variable of optimization problem is split, and the new optimization problem is solved by using Lagrangian augmented method. The experimental results show that the proposed method is faster, and the restored images have higher quality compared to the other methods.
