Loss Functions for Image Restoration With Neural Networks
TL;DR: It is shown that the quality of the results improves significantly with better loss functions, even when the network architecture is left unchanged, and a novel, differentiable error function is proposed.
Abstract: Neural networks are becoming central in several areas of computer vision and image processing and different architectures have been proposed to solve specific problems. The impact of the loss layer of neural networks, however, has not received much attention in the context of image processing: the default and virtually only choice is $\ell _2$ . In this paper, we bring attention to alternative choices for image restoration. In particular, we show the importance of perceptually-motivated losses when the resulting image is to be evaluated by a human observer. We compare the performance of several losses, and propose a novel, differentiable error function. We show that the quality of the results improves significantly with better loss functions, even when the network architecture is left unchanged.
Citations
More filters
01 Oct 2019
TL;DR: In this paper, the authors propose a set of improvements, which together result in both quantitatively and qualitatively improved depth maps compared to competing self-supervised methods, and demonstrate the effectiveness of each component in isolation, and show high quality, state-of-theart results on the KITTI benchmark.
Abstract: Per-pixel ground-truth depth data is challenging to acquire at scale. To overcome this limitation, self-supervised learning has emerged as a promising alternative for training models to perform monocular depth estimation. In this paper, we propose a set of improvements, which together result in both quantitatively and qualitatively improved depth maps compared to competing self-supervised methods. Research on self-supervised monocular training usually explores increasingly complex architectures, loss functions, and image formation models, all of which have recently helped to close the gap with fully-supervised methods. We show that a surprisingly simple model, and associated design choices, lead to superior predictions. In particular, we propose (i) a minimum reprojection loss, designed to robustly handle occlusions, (ii) a full-resolution multi-scale sampling method that reduces visual artifacts, and (iii) an auto-masking loss to ignore training pixels that violate camera motion assumptions. We demonstrate the effectiveness of each component in isolation, and show high quality, state-of-the-art results on the KITTI benchmark.
954 citations
TL;DR: A survey on recent advances of image super-resolution techniques using deep learning approaches in a systematic way, which can roughly group the existing studies of SR techniques into three major categories: supervised SR, unsupervised SR, and domain-specific SR.
Abstract: Image Super-Resolution (SR) is an important class of image processing techniqueso enhance the resolution of images and videos in computer vision. Recent years have witnessed remarkable progress of image super-resolution using deep learning techniques. This article aims to provide a comprehensive survey on recent advances of image super-resolution using deep learning approaches. In general, we can roughly group the existing studies of SR techniques into three major categories: supervised SR, unsupervised SR, and domain-specific SR. In addition, we also cover some other important issues, such as publicly available benchmark datasets and performance evaluation metrics. Finally, we conclude this survey by highlighting several future directions and open issues which should be further addressed by the community in the future.
837 citations
Cites background from "Loss Functions for Image Restoratio..."
...In practice, the L1 loss shows improved performance and convergence over L2 loss [28], [31], [126]....
[...]
Posted Content•
TL;DR: It is shown that a surprisingly simple model, and associated design choices, lead to superior predictions, and together result in both quantitatively and qualitatively improved depth maps compared to competing self-supervised methods.
Abstract: Per-pixel ground-truth depth data is challenging to acquire at scale. To overcome this limitation, self-supervised learning has emerged as a promising alternative for training models to perform monocular depth estimation. In this paper, we propose a set of improvements, which together result in both quantitatively and qualitatively improved depth maps compared to competing self-supervised methods.
Research on self-supervised monocular training usually explores increasingly complex architectures, loss functions, and image formation models, all of which have recently helped to close the gap with fully-supervised methods. We show that a surprisingly simple model, and associated design choices, lead to superior predictions. In particular, we propose (i) a minimum reprojection loss, designed to robustly handle occlusions, (ii) a full-resolution multi-scale sampling method that reduces visual artifacts, and (iii) an auto-masking loss to ignore training pixels that violate camera motion assumptions. We demonstrate the effectiveness of each component in isolation, and show high quality, state-of-the-art results on the KITTI benchmark.
791 citations
Cites methods from "Loss Functions for Image Restoratio..."
...Following [21] we use bilinear sampling to sample the source images, which is locally sub-differentiable, and we follow [75, 15] in using L1 and SSIM [64] to make our photometric error function pe, i....
[...]
18 Jun 2018
TL;DR: In this paper, a dataset of short-exposure low-light images and reference images is introduced to support the development of learning-based pipelines for low-luminance image processing.
Abstract: Imaging in low light is challenging due to low photon count and low SNR. Short-exposure images suffer from noise, while long exposure can induce blur and is often impractical. A variety of denoising, deblurring, and enhancement techniques have been proposed, but their effectiveness is limited in extreme conditions, such as video-rate imaging at night. To support the development of learning-based pipelines for low-light image processing, we introduce a dataset of raw short-exposure low-light images, with corresponding long-exposure reference images. Using the presented dataset, we develop a pipeline for processing low-light images, based on end-to-end training of a fully-convolutional network. The network operates directly on raw sensor data and replaces much of the traditional image processing pipeline, which tends to perform poorly on such data. We report promising results on the new dataset, analyze factors that affect performance, and highlight opportunities for future work.
647 citations
Posted Content•
TL;DR: A pipeline for processing low-light images is developed, based on end-to-end training of a fully-convolutional network that operates directly on raw sensor data and replaces much of the traditional image processing pipeline, which tends to perform poorly on such data.
Abstract: Imaging in low light is challenging due to low photon count and low SNR. Short-exposure images suffer from noise, while long exposure can induce blur and is often impractical. A variety of denoising, deblurring, and enhancement techniques have been proposed, but their effectiveness is limited in extreme conditions, such as video-rate imaging at night. To support the development of learning-based pipelines for low-light image processing, we introduce a dataset of raw short-exposure low-light images, with corresponding long-exposure reference images. Using the presented dataset, we develop a pipeline for processing low-light images, based on end-to-end training of a fully-convolutional network. The network operates directly on raw sensor data and replaces much of the traditional image processing pipeline, which tends to perform poorly on such data. We report promising results on the new dataset, analyze factors that affect performance, and highlight opportunities for future work. The results are shown in the supplementary video at this https URL
596 citations
Cites result from "Loss Functions for Image Restoratio..."
...Comparison in terms of PSNR/SSIM using the reference long-exposure images would not be fair to BM3D and burst processing, since these baselines have to use input images that undergo different processing....
[...]
...Table 3 (first row) reports the accuracy of the presented pipeline in terms of Peak Signal-to-Noise Ratio (PSNR) and Structural SIMilarity (SSIM) [38]....
[...]
...As shown in Table 3 (row 6), masking the Bayer data (Sony subset) yields lower PSNR/SSIM than packing; a typical perceptual artifact of the masking approach is loss of some hues in the output....
[...]
...Although images produced by the CAN have higher SSIM, they sometimes suffer from loss of color....
[...]
...As shown in Table 3 (rows 4 and 5), replacing the L1 loss by L2 or SSIM [43] produces comparable results....
[...]
References
More filters
Proceedings Article•
03 Dec 2012TL;DR: The state-of-the-art performance of CNNs was achieved by Deep Convolutional Neural Networks (DCNNs) as discussed by the authors, which consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax.
Abstract: We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes. On the test data, we achieved top-1 and top-5 error rates of 37.5% and 17.0% which is considerably better than the previous state-of-the-art. The neural network, which has 60 million parameters and 650,000 neurons, consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully-connected layers with a final 1000-way softmax. To make training faster, we used non-saturating neurons and a very efficient GPU implementation of the convolution operation. To reduce overriding in the fully-connected layers we employed a recently-developed regularization method called "dropout" that proved to be very effective. We also entered a variant of this model in the ILSVRC-2012 competition and achieved a winning top-5 test error rate of 15.3%, compared to 26.2% achieved by the second-best entry.
73,978 citations
TL;DR: In this article, a structural similarity index is proposed for image quality assessment based on the degradation of structural information, which can be applied to both subjective ratings and objective methods on a database of images compressed with JPEG and JPEG2000.
Abstract: Objective methods for assessing perceptual image quality traditionally attempted to quantify the visibility of errors (differences) between a distorted image and a reference image using a variety of known properties of the human visual system. Under the assumption that human visual perception is highly adapted for extracting structural information from a scene, we introduce an alternative complementary framework for quality assessment based on the degradation of structural information. As a specific example of this concept, we develop a structural similarity index and demonstrate its promise through a set of intuitive examples, as well as comparison to both subjective ratings and state-of-the-art objective methods on a database of images compressed with JPEG and JPEG2000. A MATLAB implementation of the proposed algorithm is available online at http://www.cns.nyu.edu//spl sim/lcv/ssim/.
40,609 citations
TL;DR: A large, deep convolutional neural network was trained to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes and employed a recently developed regularization method called "dropout" that proved to be very effective.
Abstract: We trained a large, deep convolutional neural network to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes. On the test data, we achieved top-1 and top-5 error rates of 37.5% and 17.0%, respectively, which is considerably better than the previous state-of-the-art. The neural network, which has 60 million parameters and 650,000 neurons, consists of five convolutional layers, some of which are followed by max-pooling layers, and three fully connected layers with a final 1000-way softmax. To make training faster, we used non-saturating neurons and a very efficient GPU implementation of the convolution operation. To reduce overfitting in the fully connected layers we employed a recently developed regularization method called "dropout" that proved to be very effective. We also entered a variant of this model in the ILSVRC-2012 competition and achieved a winning top-5 test error rate of 15.3%, compared to 26.2% achieved by the second-best entry.
33,301 citations
"Loss Functions for Image Restoratio..." refers background in this paper
...Following their success in several computer vision tasks [15], [2], neural networks have received considerable attention in the context of image restoration....
[...]
...Despite the promising results already produced in the 1980s on handwritten digit recognition [1], the popularity of neural networks in the field of computer vision has grown exponentially only recently, when deep learning boosted their performance in image recognition [2]....
[...]
Posted Content•
TL;DR: Caffe as discussed by the authors is a BSD-licensed C++ library with Python and MATLAB bindings for training and deploying general-purpose convolutional neural networks and other deep models efficiently on commodity architectures.
Abstract: Caffe provides multimedia scientists and practitioners with a clean and modifiable framework for state-of-the-art deep learning algorithms and a collection of reference models. The framework is a BSD-licensed C++ library with Python and MATLAB bindings for training and deploying general-purpose convolutional neural networks and other deep models efficiently on commodity architectures. Caffe fits industry and internet-scale media needs by CUDA GPU computation, processing over 40 million images a day on a single K40 or Titan GPU ($\approx$ 2.5 ms per image). By separating model representation from actual implementation, Caffe allows experimentation and seamless switching among platforms for ease of development and deployment from prototyping machines to cloud environments. Caffe is maintained and developed by the Berkeley Vision and Learning Center (BVLC) with the help of an active community of contributors on GitHub. It powers ongoing research projects, large-scale industrial applications, and startup prototypes in vision, speech, and multimedia.
12,531 citations
Posted Content•
TL;DR: This work proposes a Parametric Rectified Linear Unit (PReLU) that generalizes the traditional rectified unit and derives a robust initialization method that particularly considers the rectifier nonlinearities.
Abstract: Rectified activation units (rectifiers) are essential for state-of-the-art neural networks. In this work, we study rectifier neural networks for image classification from two aspects. First, we propose a Parametric Rectified Linear Unit (PReLU) that generalizes the traditional rectified unit. PReLU improves model fitting with nearly zero extra computational cost and little overfitting risk. Second, we derive a robust initialization method that particularly considers the rectifier nonlinearities. This method enables us to train extremely deep rectified models directly from scratch and to investigate deeper or wider network architectures. Based on our PReLU networks (PReLU-nets), we achieve 4.94% top-5 test error on the ImageNet 2012 classification dataset. This is a 26% relative improvement over the ILSVRC 2014 winner (GoogLeNet, 6.66%). To our knowledge, our result is the first to surpass human-level performance (5.1%, Russakovsky et al.) on this visual recognition challenge.
11,866 citations
"Loss Functions for Image Restoratio..." refers background or methods in this paper
...Following their success in several computer vision tasks [15], [2], neural networks have received considerable attention in the context of image restoration....
[...]
...We apply parametric rectified linear unit (PReLU) layers to the output of the inner convolutional layers [15]....
[...]