PyTorchKR Papers

최신 논문

총 3,696편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2011.01748공식 코드 · Papers with Code

Solving Inverse Problems with Hybrid Deep Image Priors: the challenge of preventing overfitting

We mainly analyze and solve the overfitting problem of deep image prior (DIP). Deep image prior can solve inverse problems such as super-resolution, inpainting and denoising. The main advantage of DIP over other deep learning approaches is that it does not need access to a large dataset. However, due to the large number of parameters of the neural network and noisy data, DIP overfits to the noise in the image as the number of iterations grows. In the thesis, we use hybrid deep image priors to avoid overfitting. The hybrid priors are to combine DIP with an explicit prior such as total variation or with an implicit prior such as a denoising algorithm. We use the alternating direction method-of-multipliers (ADMM) to incorporate the new prior and try different forms of ADMM to avoid extra computation caused by the inner loop of ADMM steps. We also study the relation between the dynamics of gradient descent, and the overfitting phenomenon. The numerical results show the hybrid priors play an important role in preventing overfitting. Besides, we try to fit the image along some directions and find this method can reduce overfitting when the noise level is large. When the noise level is small, it does not considerably reduce the overfitting problem.

arXiv:2011.01741

Learning a Generative Motion Model from Image Sequences based on a Latent Motion Matrix

We propose to learn a probabilistic motion model from a sequence of images for spatio-temporal registration. Our model encodes motion in a low-dimensional probabilistic space - the motion matrix - which enables various motion analysis tasks such as simulation and interpolation of realistic motion patterns allowing for faster data acquisition and data augmentation. More precisely, the motion matrix allows to transport the recovered motion from one subject to another simulating for example a pathological motion in a healthy subject without the need for inter-subject registration. The method is based on a conditional latent variable model that is trained using amortized variational inference. This unsupervised generative model follows a novel multivariate Gaussian process prior and is applied within a temporal convolutional network which leads to a diffeomorphic motion model. Temporal consistency and generalizability is further improved by applying a temporal dropout training scheme. Applied to cardiac cine-MRI sequences, we show improved registration accuracy and spatio-temporally smoother deformations compared to three state-of-the-art registration algorithms. Besides, we demonstrate the model's applicability for motion analysis, simulation and super-resolution by an improved motion reconstruction from sequences with missing frames compared to linear and cubic interpolation.

arXiv:2011.01926

Generating Unobserved Alternatives

We consider problems where multiple predictions can be considered correct, but only one of them is given as supervision. This setting differs from both the regression and class-conditional generative modelling settings: in the former, there is a unique observed output for each input, which is provided as supervision; in the latter, there are many observed outputs for each input, and many are provided as supervision. Applying either regression methods and conditional generative models to the present setting often results in a model that can only make a single prediction for each input. We explore several problems that have this property and develop an approach that can generate multiple high-quality predictions given the same input. As a result, it can be used to generate high-quality outputs that are different from the observed output.

arXiv:2010.15491

A Novel Fast 3D Single Image Super-Resolution Algorithm

This paper introduces a novel computationally efficient method of solving the 3D single image super-resolution (SR) problem, i.e., reconstruction of a high-resolution volume from its low-resolution counterpart. The main contribution lies in the original way of handling simultaneously the associated decimation and blurring operators, based on their underlying properties in the frequency domain. In particular, the proposed decomposition technique of the 3D decimation operator allows a straightforward implementation for Tikhonov regularization, and can be further used to take into consideration other regularization functions such as the total variation, enabling the computational cost of state-of-the-art algorithms to be considerably decreased. Numerical experiments carried out showed that the proposed approach outperforms existing 3D SR methods.

arXiv:2010.15689공식 코드 · Papers with Code

Learning Deep Interleaved Networks with Asymmetric Co-Attention for Image Restoration

Recently, convolutional neural network (CNN) has demonstrated significant success for image restoration (IR) tasks (e.g., image super-resolution, image deblurring, rain streak removal, and dehazing). However, existing CNN based models are commonly implemented as a single-path stream to enrich feature representations from low-quality (LQ) input space for final predictions, which fail to fully incorporate preceding low-level contexts into later high-level features within networks, thereby producing inferior results. In this paper, we present a deep interleaved network (DIN) that learns how information at different states should be combined for high-quality (HQ) images reconstruction. The proposed DIN follows a multi-path and multi-branch pattern allowing multiple interconnected branches to interleave and fuse at different states. In this way, the shallow information can guide deep representative features prediction to enhance the feature expression ability. Furthermore, we propose asymmetric co-attention (AsyCA) which is attached at each interleaved node to model the feature dependencies. Such AsyCA can not only adaptively emphasize the informative features from different states, but also improves the discriminative ability of networks. Our presented DIN can be trained end-to-end and applied to various IR tasks. Comprehensive evaluations on public benchmarks and real-world datasets demonstrate that the proposed DIN perform favorably against the state-of-the-art methods quantitatively and qualitatively.

arXiv:2010.14714

Differentiable Channel Sparsity Search via Weight Sharing within Filters

In this paper, we propose the differentiable channel sparsity search (DCSS) for convolutional neural networks. Unlike traditional channel pruning algorithms which require users to manually set prune ratios for each convolutional layer, DCSS automatically searches the optimal combination of sparsities. Inspired by the differentiable architecture search (DARTS), we draw lessons from the continuous relaxation and leverage the gradient information to balance the computational cost and metrics. Since directly applying the scheme of DARTS causes shape mismatching and excessive memory consumption, we introduce a novel technique called weight sharing within filters. This technique elegantly eliminates the problem of shape mismatching with negligible additional resources. We conduct comprehensive experiments on not only image classification but also find-grained tasks including semantic segmentation and image super resolution to verify the effectiveness of DCSS. Compared with previous network pruning approaches, DCSS achieves state-of-the-art results for image classification. Experimental results of semantic segmentation and image super resolution indicate that task-specific search achieves better performance than transferring slim models, demonstrating the wide applicability and high efficiency of DCSS.

arXiv:2010.12678

Super-Resolution Reconstruction of Interval Energy Data

High-resolution data are desired in many data-driven applications; however, in many cases only data whose resolution is lower than expected are available due to various reasons. It is then a challenge how to obtain as much useful information as possible from the low-resolution data. In this paper, we target interval energy data collected by Advanced Metering Infrastructure (AMI), and propose a Super-Resolution Reconstruction (SRR) approach to upsample low-resolution (hourly) interval data into higher-resolution (15-minute) data using deep learning. Our preliminary results show that the proposed SRR approaches can achieve much improved performance compared to the baseline model.

arXiv:2010.13172

Unsupervised Super-Resolution: Creating High-Resolution Medical Images from Low-Resolution Anisotropic Examples

Although high resolution isotropic 3D medical images are desired in clinical practice, their acquisition is not always feasible. Instead, lower resolution images are upsampled to higher resolution using conventional interpolation methods. Sophisticated learning-based super-resolution approaches are frequently unavailable in clinical setting, because such methods require training with high-resolution isotropic examples. To address this issue, we propose a learning-based super-resolution approach that can be trained using solely anisotropic images, i.e. without high-resolution ground truth data. The method exploits the latent space, generated by autoencoders trained on anisotropic images, to increase spatial resolution in low-resolution images. The method was trained and evaluated using 100 publicly available cardiac cine MR scans from the Automated Cardiac Diagnosis Challenge (ACDC). The quantitative results show that the proposed method performs better than conventional interpolation methods. Furthermore, the qualitative results indicate that especially finer cardiac structures are synthesized with high quality. The method has the potential to be applied to other anatomies and modalities and can be easily applied to any 3D anisotropic medical image dataset.

arXiv:2010.13821

Wavelet Flow: Fast Training of High Resolution Normalizing Flows

Normalizing flows are a class of probabilistic generative models which allow for both fast density computation and efficient sampling and are effective at modelling complex distributions like images. A drawback among current methods is their significant training cost, sometimes requiring months of GPU training time to achieve state-of-the-art results. This paper introduces Wavelet Flow, a multi-scale, normalizing flow architecture based on wavelets. A Wavelet Flow has an explicit representation of signal scale that inherently includes models of lower resolution signals and conditional generation of higher resolution signals, i.e., super resolution. A major advantage of Wavelet Flow is the ability to construct generative models for high resolution data (e.g., 1024 x 1024 images) that are impractical with previous models. Furthermore, Wavelet Flow is competitive with previous normalizing flows in terms of bits per dimension on standard (low resolution) benchmarks while being up to 15x faster to train.

arXiv:2010.14105

Micro-CT Synthesis and Inner Ear Super Resolution via Generative Adversarial Networks and Bayesian Inference

Existing medical image super-resolution methods rely on pairs of low- and high- resolution images to learn a mapping in a fully supervised manner. However, such image pairs are often not available in clinical practice. In this paper, we address super-resolution problem in a real-world scenario using unpaired data and synthesize linearly \textbf{eight times} higher resolved Micro-CT images of temporal bone structure, which is embedded in the inner ear. We explore cycle-consistency generative adversarial networks for super-resolution task and equip the translation approach with Bayesian inference. We further introduce \emph{Hu Moment distance} the evaluation metric to quantify the shape of the temporal bone. We evaluate our method on a public inner ear CT dataset and have seen both visual and quantitative improvement over state-of-the-art deep-learning-based methods. In addition, we perform a multi-rater visual evaluation experiment and find that trained experts consistently rate the proposed method the highest quality scores among all methods. Furthermore, we are able to quantify uncertainty in the unpaired translation task and the uncertainty map can provide structural information of the temporal bone.

arXiv:2010.12576

Non-convex Super-resolution of OCT images via sparse representation

We propose a non-convex variational model for the super-resolution of Optical Coherence Tomography (OCT) images of the murine eye, by enforcing sparsity with respect to suitable dictionaries learnt from high-resolution OCT data. The statistical characteristics of OCT images motivate the use of {\alpha}-stable distributions for learning dictionaries, by considering the non-Gaussian case, {\alpha}=1. The sparsity-promoting cost function relies on a non-convex penalty - Cauchy-based or Minimax Concave Penalty (MCP) - which makes the problem particularly challenging. We propose an efficient algorithm for minimizing the function based on the forward-backward splitting strategy which guarantees at each iteration the existence and uniqueness of the proximal point. Comparisons with standard convex L1-based reconstructions show the better performance of non-convex models, especially in view of further OCT image analysis

arXiv:2010.12249공식 코드 · Papers with Code

Multi Scale Identity-Preserving Image-to-Image Translation Network for Low-Resolution Face Recognition

State-of-the-art deep neural network models have reached near perfect face recognition accuracy rates on controlled high-resolution face images. However, their performance is drastically degraded when they are tested with very low-resolution face images. This is particularly critical in surveillance systems, where a low-resolution probe image is to be matched with high-resolution gallery images. super-resolution techniques aim at producing high-resolution face images from low-resolution counterparts. While they are capable of reconstructing images that are visually appealing, the identity-related information is not preserved. Here, we propose an identity-preserving end-to-end image-to-image translation deep neural network which is capable of super-resolving very low-resolution faces to their high-resolution counterparts while preserving identity-related information. We achieved this by training a very deep convolutional encoder-decoder network with a symmetric contracting path between corresponding layers. This network was trained with a combination of a reconstruction and an identity-preserving loss, on multi-scale low-resolution conditions. Extensive quantitative evaluations of our proposed model demonstrated that it outperforms competing super-resolution and low-resolution face recognition methods on natural and artificial low-resolution face data sets and even unseen identities.

arXiv:2010.11348

Deep Learning for Efficient Reconstruction of High-Resolution Turbulent DNS Data

Within the domain of Computational Fluid Dynamics, Direct Numerical Simulation (DNS) is used to obtain highly accurate numerical solutions for fluid flows. However, this approach for numerically solving the Navier-Stokes equations is extremely computationally expensive mostly due to the requirement of greatly refined grids. Large Eddy Simulation (LES) presents a more computationally efficient approach for solving fluid flows on lower-resolution (LR) grids but results in an overall reduction in solution fidelity. Through this paper, we introduce a novel deep learning framework SR-DNS Net, which aims to mitigate this inherent trade-off between solution fidelity and computational complexity by leveraging deep learning techniques used in image super-resolution. Using our model, we wish to learn the mapping from a coarser LR solution to a refined high-resolution (HR) DNS solution so as to eliminate the need for performing DNS on highly refined grids. Our model efficiently reconstructs the high-fidelity DNS data from the LES like low-resolution solutions while yielding good reconstruction metrics. Thus our implementation improves the solution accuracy of LR solutions while incurring only a marginal increase in computational cost required for deploying the trained deep learning model.

arXiv:2010.11575공식 코드 · Papers with Code

Face Hallucination via Split-Attention in Split-Attention Network

Recently, convolutional neural networks (CNNs) have been widely employed to promote the face hallucination due to the ability to predict high-frequency details from a large number of samples. However, most of them fail to take into account the overall facial profile and fine texture details simultaneously, resulting in reduced naturalness and fidelity of the reconstructed face, and further impairing the performance of downstream tasks (e.g., face detection, facial recognition). To tackle this issue, we propose a novel external-internal split attention group (ESAG), which encompasses two paths responsible for facial structure information and facial texture details, respectively. By fusing the features from these two paths, the consistency of facial structure and the fidelity of facial details are strengthened at the same time. Then, we propose a split-attention in split-attention network (SISN) to reconstruct photorealistic high-resolution facial images by cascading several ESAGs. Experimental results on face hallucination and face recognition unveil that the proposed method not only significantly improves the clarity of hallucinated faces, but also encourages the subsequent face recognition performance substantially. Codes have been released at https://github.com/mdswyz/SISN-Face-Hallucination.

arXiv:2010.11083

Adaptive Pixel-wise Structured Sparse Network for Efficient CNNs

To accelerate deep CNN models, this paper proposes a novel spatially adaptive framework that can dynamically generate pixel-wise sparsity according to the input image. The sparse scheme is pixel-wise refined, regional adaptive under a unified importance map, which makes it friendly to hardware implementation. A sparse controlling method is further presented to enable online adjustment for applications with different precision/latency requirements. The sparse model is applicable to a wide range of vision tasks. Experimental results show that this method efficiently improve the computing efficiency for both image classification using ResNet-18 and super resolution using SRResNet. On image classification task, our method can save 30%-70% MACs with a slightly drop in top-1 and top-5 accuracy. On super resolution task, our method can reduce more than 90% MACs while only causing around 0.1 dB and 0.01 decreasing in PSNR and SSIM. Hardware validation is also included.

arXiv:2010.10207

Micro CT Image-Assisted Cross Modality Super-Resolution of Clinical CT Images Utilizing Synthesized Training Dataset

This paper proposes a novel, unsupervised super-resolution (SR) approach for performing the SR of a clinical CT into the resolution level of a micro CT ($\mu$CT). The precise non-invasive diagnosis of lung cancer typically utilizes clinical CT data. Due to the resolution limitations of clinical CT (about $0.5 \times 0.5 \times 0.5$ mm$^3$), it is difficult to obtain enough pathological information such as the invasion area at alveoli level. On the other hand, $\mu$CT scanning allows the acquisition of volumes of lung specimens with much higher resolution ($50 \times 50 \times 50 \mu {\rm m}^3$ or higher). Thus, super-resolution of clinical CT volume may be helpful for diagnosis of lung cancer. Typical SR methods require aligned pairs of low-resolution (LR) and high-resolution (HR) images for training. Unfortunately, obtaining paired clinical CT and $\mu$CT volumes of human lung tissues is infeasible. Unsupervised SR methods are required that do not need paired LR and HR images. In this paper, we create corresponding clinical CT-$\mu$CT pairs by simulating clinical CT images from $\mu$CT images by modified CycleGAN. After this, we use simulated clinical CT-$\mu$CT image pairs to train an SR network based on SRGAN. Finally, we use the trained SR network to perform SR of the clinical CT images. We compare our proposed method with another unsupervised SR method for clinical CT images named SR-CycleGAN. Experimental results demonstrate that the proposed method can successfully perform SR of clinical CT images of lung cancer patients with $\mu$CT level resolution, and quantitatively and qualitatively outperformed conventional method (SR-CycleGAN), improving the SSIM (structure similarity) form 0.40 to 0.51.

arXiv:2010.08919

Boosting High-Level Vision with Joint Compression Artifacts Reduction and Super-Resolution

Due to the limits of bandwidth and storage space, digital images are usually down-scaled and compressed when transmitted over networks, resulting in loss of details and jarring artifacts that can lower the performance of high-level visual tasks. In this paper, we aim to generate an artifact-free high-resolution image from a low-resolution one compressed with an arbitrary quality factor by exploring joint compression artifacts reduction (CAR) and super-resolution (SR) tasks. First, we propose a context-aware joint CAR and SR neural network (CAJNN) that integrates both local and non-local features to solve CAR and SR in one-stage. Finally, a deep reconstruction network is adopted to predict high quality and high-resolution images. Evaluation on CAR and SR benchmark datasets shows that our CAJNN model outperforms previous methods and also takes 26.2% shorter runtime. Based on this model, we explore addressing two critical challenges in high-level computer vision: optical character recognition of low-resolution texts, and extremely tiny face detection. We demonstrate that CAJNN can serve as an effective image preprocessing method and improve the accuracy for real-scene text recognition (from 85.30% to 85.75%) and the average precision for tiny face detection (from 0.317 to 0.611).

arXiv:2010.08895공식 코드 · Papers with Code

Fourier Neural Operator for Parametric Partial Differential Equations

The classical development of neural networks has primarily focused on learning mappings between finite-dimensional Euclidean spaces. Recently, this has been generalized to neural operators that learn mappings between function spaces. For partial differential equations (PDEs), neural operators directly learn the mapping from any functional parametric dependence to the solution. Thus, they learn an entire family of PDEs, in contrast to classical methods which solve one instance of the equation. In this work, we formulate a new neural operator by parameterizing the integral kernel directly in Fourier space, allowing for an expressive and efficient architecture. We perform experiments on Burgers' equation, Darcy flow, and Navier-Stokes equation. The Fourier neural operator is the first ML-based method to successfully model turbulent flows with zero-shot super-resolution. It is up to three orders of magnitude faster compared to traditional PDE solvers. Additionally, it achieves superior accuracy compared to previous learning-based solvers under fixed resolution.

arXiv:2010.08888

Light Stage Super-Resolution: Continuous High-Frequency Relighting

The light stage has been widely used in computer graphics for the past two decades, primarily to enable the relighting of human faces. By capturing the appearance of the human subject under different light sources, one obtains the light transport matrix of that subject, which enables image-based relighting in novel environments. However, due to the finite number of lights in the stage, the light transport matrix only represents a sparse sampling on the entire sphere. As a consequence, relighting the subject with a point light or a directional source that does not coincide exactly with one of the lights in the stage requires interpolation and resampling the images corresponding to nearby lights, and this leads to ghosting shadows, aliased specularities, and other artifacts. To ameliorate these artifacts and produce better results under arbitrary high-frequency lighting, this paper proposes a learning-based solution for the "super-resolution" of scans of human faces taken from a light stage. Given an arbitrary "query" light direction, our method aggregates the captured images corresponding to neighboring lights in the stage, and uses a neural network to synthesize a rendering of the face that appears to be illuminated by a "virtual" light source at the query location. This neural network must circumvent the inherent aliasing and regularity of the light stage data that was used for training, which we accomplish through the use of regularized traditional interpolation methods within our network. Our learned model is able to produce renderings for arbitrary light directions that exhibit realistic shadows and specular highlights, and is able to generalize across a wide variety of subjects.

arXiv:2010.09594

Multi-Modal Super Resolution for Dense Microscopic Particle Size Estimation

Particle Size Analysis (PSA) is an important process carried out in a number of industries, which can significantly influence the properties of the final product. A ubiquitous instrument for this purpose is the Optical Microscope (OM). However, OMs are often prone to drawbacks like low resolution, small focal depth, and edge features being masked due to diffraction. We propose a powerful application of a combination of two Conditional Generative Adversarial Networks (cGANs) that Super Resolve OM images to look like Scanning Electron Microscope (SEM) images. We further demonstrate the use of a custom object detection module that can perform efficient PSA of the super-resolved particles on both, densely and sparsely packed images. The PSA results obtained from the super-resolved images have been benchmarked against human annotators, and results obtained from the corresponding SEM images. The proposed models show a generalizable way of multi-modal image translation and super-resolution for accurate particle size estimation.