Single Image Super-resolution Models Research Articles

Denoising Diffusion Probabilistic Model with Adversarial Learning for Remote Sensing Super-Resolution

Single Image Super-Resolution (SISR) for image enhancement enables the generation of high spatial resolution in Remote Sensing (RS) images without incurring additional costs. This approach offers a practical solution to obtain high-resolution RS images, addressing challenges posed by the expense of acquisition equipment and unpredictable weather conditions. To address the over-smoothing of the previous SISR models, the diffusion model has been incorporated into RS SISR to generate Super-Resolution (SR) images with enhanced textural details. In this paper, we propose a Diffusion model with Adversarial Learning Strategy (DiffALS) to refine the generative capability of the diffusion model. DiffALS integrates an additional Noise Discriminator (ND) into the training process, employing an adversarial learning strategy on the data distribution learning. This ND guides noise prediction by considering the general correspondence between the noisy image in each step, thereby enhancing the diversity of generated data and the detailed texture prediction of the diffusion model. Furthermore, considering that the diffusion model may exhibit suboptimal performance on traditional pixel-level metrics such as Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity (SSIM), we showcase the effectiveness of DiffALS through downstream semantic segmentation applications. Extensive experiments demonstrate that the proposed model achieves remarkable accuracy and notable visual enhancements. Compared to other state-of-the-art methods, our model establishes an improvement of 189 for Fréchet Inception Distance (FID) and 0.002 for Learned Perceptual Image Patch Similarity (LPIPS) in a SR dataset, namely Alsat, and achieves improvements of 0.4%, 0.3%, and 0.2% for F1 score, MIoU, and Accuracy, respectively, in a segmentation dataset, namely Vaihingen.

A Multi-Attention Feature Distillation Neural Network for Lightweight Single Image Super-Resolution

In recent years, remarkable performance improvements have been produced by deep convolutional neural networks (CNN) for single image super-resolution (SISR). Nevertheless, a high proportion of CNN-based SISR models are with quite a few network parameters and high computational complexity for deep or wide architectures. How to more fully utilize deep features to make a balance between model complexity and reconstruction performance is one of the main challenges in this field. To address this problem, on the basis of the well-known information multi-distillation model, a multi-attention feature distillation network termed as MAFDN is developed for lightweight and accurate SISR. Specifically, an effective multi-attention feature distillation block (MAFDB) is designed and used as the basic feature extraction unit in MAFDN. With the help of multi-attention layers including pixel attention, spatial attention, and channel attention, MAFDB uses multiple information distillation branches to learn more discriminative and representative features. Furthermore, MAFDB introduces the depthwise over-parameterized convolutional layer (DO-Conv)-based residual block (OPCRB) to enhance its ability without incurring any parameter and computation increase in the inference stage. The results on commonly used datasets demonstrate that our MAFDN outperforms existing representative lightweight SISR models when taking both reconstruction performance and model complexity into consideration. For example, for ×4 SR on Set5, MAFDN (597K/33.79G) obtains 0.21 dB/0.0037 and 0.10 dB/0.0015 PSNR/SSIM gains over the attention-based SR model AFAN (692K/50.90G) and the feature distillation-based SR model DDistill-SR (675K/32.83G), respectively.

RefQSR: Reference-Based Quantization for Image Super-Resolution Networks.

Single image super-resolution (SISR) aims to reconstruct a high-resolution image from its low-resolution observation. Recent deep learning-based SISR models show high performance at the expense of increased computational costs, limiting their use in resource-constrained environments. As a promising solution for computationally efficient network design, network quantization has been extensively studied. However, existing quantization methods developed for SISR have yet to effectively exploit image self-similarity, which is a new direction for exploration in this study. We introduce a novel method called reference-based quantization for image super-resolution (RefQSR) that applies high-bit quantization to several representative patches and uses them as references for low-bit quantization of the rest of the patches in an image. To this end, we design dedicated patch clustering and reference-based quantization modules and integrate them into existing SISR network quantization methods. The experimental results demonstrate the effectiveness of RefQSR on various SISR networks and quantization methods.

Two-Stage Deep Single-Image Super-Resolution With Multiple Blur Kernels for Internet of Things

Single image super-resolution (SISR) aims to reconstruct a high-resolution image from a single low-resolution (LR) image. Although convolutional neural network (CNN)-based SISR methods greatly enhance image restoration, they face critical challenges. First, SISR models using CNNs process image patches uniformly regardless of importance, causing spatial inefficiency in computation and representation. However, due to resource constraints, edge devices in the Internet of Things (IoT) cannot bear heavy computations or large memory storage. Second, most of the existing SISR methods are designed only for the widely adopted bicubic degradation and cannot handle LR images with arbitrary blur kernels, resulting in poor recovery performance. To address these issues, in this paper we propose a two-stage semantic and spatial deep super-resolution (SSDSR) model suitable for the IoT environment. The proposed SSDSR model is capable of handling a variety of blur kernels (e.g., isotropic Gaussian, motion and disk blur) by effectively using their prior information. Moreover, the Semantic Feature Extraction (SFE) module enables the proposed model to focus on key areas of LR images rather than treating all pixels equally, which significantly reduces the computational load. The semantic information from the SFE module and the spatial information from the Spatial Attention module are fused adaptively, allowing the proposed model to extract key information in LR images, thereby increasing the representation capacity of the CNN and improving image recovery. Compared with state-of-the-art SISR methods on benchmark datasets, the proposed SSDSR model demonstrates superior performance. When run in real time on an IoT edge device, our model exhibits high computational efficiency and excellent image quality.

Improved wavelet prediction superresolution reconstruction based on U‐Net

AbstractDeep learning can be used to achieve single‐image superresolution (SR) reconstruction. To address problems encountered during this process, such as the number of network parameters, high training requirements on equipment performance, and inability to downsample certain SR images accurately, an image SR reconstruction algorithm based on deep residual network optimization is proposed. The model introduces wavelet transforms based on the original U‐Net, where the U‐Net is trained to obtain SR wavelet feature images at multiple scales simultaneously. This approach reduces the mapping space for the network to learn low‐ to high‐resolution image mapping, which in turn reduces the training difficulty of the model. In terms of network details, the inverse wavelet transform is used in image upsampling to enhance the sparsity of the reconstruction layer in the original network. The network structure of the U‐Net upsampling is adjusted slightly to enable the network to distinguish wavelet images from feature images, thereby improving the richness of the features extracted by the model. The experimental results show that the peak signal‐to‐noise ratio (PSNR) of the fourfold SR model is 32.35 and 28.68 on the Set5 and Set14 validation sets, respectively. Compared with networks that use wavelet prediction mechanisms, such as the deep wavelet prediction SR (DWSR) and deep wavelet prediction‐based residual SR (DWRSR) models, the PSNR for all the tested public datasets is improved by 0.5. The method yields superior results in terms of both visual effect and PSNR, demonstrating the feasibility of the wavelet prediction mechanism in SR reconstruction and thus offering application value and research significance.

Fast Non-Local Attention network for light super-resolution

Although convolutional neural network-based methods have achieved significant performance improvement for Single Image Super-Resolution (SISR), their vast computational cost hinders real-world environment application. Thus, the interest in light networks for SISR is rising. Since existing SISR light models mainly focus on extracting fine local features using convolution operation, they have a limitation in that networks hardly capture global information. To capture the long-range dependency, Non-Local (NL) attention and Transformers have been explored in the SISR task. However, they are still suffering from a balancing problem between performance and computational cost. In this paper, we propose Fast Non-Local attention NETwork (FNLNET) for a super light SISR, which can capture the global representation. To acquire global information, we propose The Fast Non-Local Attention (FNLA) module that has low computational complexity while capturing global representation that reflects long-distance relationships between patches. Then, FNLA requires only 16 times lower computational cost than conventional NL networks while improving performance. In addition, we propose a powerful module called Global Self-Intension Mining (GSIM) that fuses the multi-information resources such as local, and global representation. Our FNLNET shows outstanding performance with fewer parameters and computational costs in the experiments on the benchmark datasets against state-of-the-art light SISR models.

Differentiable Neural Architecture Search for Extremely Lightweight Image Super-Resolution

Single Image Super-Resolution (SISR) tasks have achieved significant performance with deep neural networks. However, the large number of parameters in CNN-based methods for SISR tasks require heavy computations. Although several efficient SISR models have been recently proposed, most are handcrafted and thus lack flexibility. In this work, we propose a novel differentiable Neural Architecture Search (NAS) approach on both the cell-level and network-level to search for lightweight SISR models. Specifically, the cell-level search space is designed based on an information distillation mechanism, focusing on the combinations of lightweight operations and aiming to build a more lightweight and accurate SR structure. The network-level search space is designed to consider the feature connections among the cells and aims to find which information flow benefits the cell most to boost the performance. Unlike the existing Reinforcement Learning (RL) or Evolutionary Algorithm (EA) based NAS methods for SISR tasks, our search pipeline is fully differentiable, and the lightweight SISR models can be efficiently searched on both the cell-level and network-level jointly on a single GPU. Experiments show that our methods can achieve state-of-the-art performance on the benchmark datasets in terms of PSNR, SSIM, and model complexity with merely 68G Multi-Adds for ×2 and 18G Multi-Adds for ×4 SR tasks.

Adapting Single-Image Super-Resolution Models to Video Super-Resolution: A Plug-and-Play Approach.

The quality of videos varies due to the different capabilities of sensors. Video super-resolution (VSR) is a technology that improves the quality of captured video. However, the development of a VSR model is very costly. In this paper, we present a novel approach for adapting single-image super-resolution (SISR) models to the VSR task. To achieve this, we first summarize a common architecture of SISR models and perform a formal analysis of adaptation. Then, we propose an adaptation method that incorporates a plug-and-play temporal feature extraction module into existing SISR models. The proposed temporal feature extraction module consists of three submodules: offset estimation, spatial aggregation, and temporal aggregation. In the spatial aggregation submodule, the features obtained from the SISR model are aligned to the center frame based on the offset estimation results. The aligned features are fused in the temporal aggregation submodule. Finally, the fused temporal feature is fed to the SISR model for reconstruction. To evaluate the effectiveness of our method, we adapt five representative SISR models and evaluate these models on two popular benchmarks. The experiment results show the proposed method is effective on different SISR models. In particular, on the Vid4 benchmark, the VSR-adapted models achieve at least 1.26 dB and 0.067 improvement over the original SISR models in terms of PSNR and SSIM metrics, respectively. Additionally, these VSR-adapted models achieve better performance than the state-of-the-art VSR models.

Multiregression spatially variant blur kernel estimation based on inter‐kernel consistency

AbstractMost single‐image super‐resolution (SR) models suffer from the degradation of image restoration performance when restoring a high‐resolution (HR) image from a low‐resolution (LR) image downscaled using an unknown blur kernel. The spatially invariant blur kernel estimators have been proposed to predict the blur kernel to address this issue. Nevertheless, the spatially variant blur exists in the real‐world; thus, these blur kernel estimators are unsuitable for real‐world applications. Although the spatially variant blur kernel estimators have been proposed, SR models still suffer from performance degradation because these estimators do neither consider the consistency between surrounding blur kernels nor refine non‐parametric blur kernels as parameters. To address this problem, the authors propose a multiregression spatially variant blur kernel estimation based on inter‐kernel consistency. The proposed estimator consists of three parts: non‐parametric regression, an inter‐kernel consistency block, and parametric regression. Specifically, it predicts spatially variant blur kernels while considering the inter‐kernel consistency between nearby blur kernels. Our source codes with pretrained models are available on https://github.com/alsgur0720/multiregression.

Global Learnable Attention for Single Image Super-Resolution.

Self-similarity is valuable to the exploration of non-local textures in single image super-resolution (SISR). Researchers usually assume that the importance of non-local textures is positively related to their similarity scores. In this paper, we surprisingly found that when repairing severely damaged query textures, some non-local textures with low-similarity which are closer to the target can provide more accurate and richer details than the high-similarity ones. In these cases, low-similarity does not mean inferior but is usually caused by different scales or orientations. Utilizing this finding, we proposed a Global Learnable Attention (GLA) to adaptively modify similarity scores of non-local textures during training instead of only using a fixed similarity scoring function such as the dot product. The proposed GLA can explore non-local textures with low-similarity but more accurate details to repair severely damaged textures. Furthermore, we propose to adopt Super-Bit Locality-Sensitive Hashing (SB-LSH) as a preprocessing method for our GLA. With the SB-LSH, the computational complexity of our GLA is reduced from quadratic to asymptotic linear with respect to the image size. In addition, the proposed GLA can be integrated into existing deep SISR models as an efficient general building block. Based on the GLA, we constructed a Deep Learnable Similarity Network (DLSN), which achieves state-of-the-art performance for SISR tasks of different degradation types (e.g., blur and noise). Our code and a pre-trained DLSN have been uploaded to GitHub† for validation.

A Data Enhancement Method toward Real-World Single Image Super-Resolution

Introducing convolutional neural networks (CNNs) into single image super-resolution (SISR) has achieved remarkable performance in experimental scenarios over the past decades. However, maintaining constant mapping complexity, most existing CNN-based SISR approaches make it challenging to reconstruct high-quality images in real scenarios because of considerable variation between authentic degradation and simulated degradation. In this paper, to improve SISR models’ capabilities in the real world, we propose a data enhancement method, consisting of frequency domain data augmentation (FDA) and dynamical frequency loss (DFL). By introducing high-resolution images’ frequency information, the FDA enhances the quality and quantity of data, regularizing the model to learn “how” to appropriately apply super-resolution on different frequency components of a given image. In addition, to further strengthen the regularization effect, we design DFL to dynamically assign the loss trade-off hyperparameters of distinct frequency components, guiding the model to allocate computational overhead reasonably. Considering quantitative analysis and visual quality in combination, extensive experiments prove that the proposed enhancement method significantly improves SR models’ capabilities in real scenarios without adding any computational cost to the evaluation procedure.

Single image super‐resolution based on progressive fusion of orientation‐aware features

Single image super-resolution (SISR) is an active research topic in the fields of image processing, computer vision and pattern recognition, restoring high-frequency details and textures based on the low-resolution input image. In this paper, we aim to build more accurate and faster SISR models via developing better-performing feature extraction and fusion techniques. Firstly, we proposed a novel Orientation-Aware feature extraction/selection Module (OAM), which contains a mixture of 1D and 2D convolutional kernels (i.e., 3×1, 1×3, and 3×3) for extracting orientation-aware features. The channel attention mechanism is deployed within each OAM, performing scene-specific selection of informative outputs of the orientation-dependent kernels (e.g., horizontal, vertical, and diagonal). Secondly, we present an effective fusion architecture to progressively integrate multi-scale features extracted in different convolutional stages. Instead of directly combining low-level and high-level features, similar outputs of adjacent feature extraction modules are grouped and further compressed to generate a more concise representation of a specific convolutional stage for high-accuracy SISR task. Based on the above two important improvements, we present a compact but effective CNN-based model for high-quality SISR via Progressive Fusion of Orientation-Aware features (SISR-PF-OA). Extensive experimental results verify the superiority of the proposed SISR-PF-OA model, performing favorably against the state-of-the-art models in terms of both restoration accuracy and computational efficiency (e.g., SISR-PF-OA outperforms RCAN model, achieving higher PSNR 31.25 dB vs. 31.21 dB and using fewer FLOPs 764.41 G vs. 1020.28 G on the Manga109 dataset for scale factor ×4 SISR task.). The source codes will be made publicly available.

Self-supervised cycle-consistent learning for scale-arbitrary real-world single image super-resolution

Whether conventional machine learning-based or current deep neural networks-based single image super-resolution (SISR) methods, they are generally trained and validated on synthetic datasets, in which low-resolution (LR) inputs are artificially produced by degrading high-resolution (HR) images based on a hand-crafted degradation model (e.g., bicubic downsampling). One of the main reasons for this is that it is challenging to build a realistic dataset composed of real-world LR–HR image pairs. However, a domain gap exists between synthetic and real-world data because the degradations in real scenarios are more complicated, limiting the performance in practical applications of SISR models trained with synthetic data. To address these problems, we propose a Self-supervised Cycle-consistent Learning-based Scale-Arbitrary Super-Resolution framework (SCL-SASR) for real-world images. Inspired by the Maximum a Posteriori estimation, our SCL-SASR consists of a Scale-Arbitrary Super-Resolution Network (SASRN) and an inverse Scale-Arbitrary Resolution-Degradation Network (SARDN). SARDN and SASRN restrain each other with the bidirectional cycle consistency constraints as well as image priors, making SASRN adapt to the image-specific degradation well. Meanwhile, considering the lack of targeted training images and the complexity of realistic degradations, SCL-SASR is designed to be online optimized solely with the LR input prior to the SR reconstruction. Benefitting from the flexible architecture and the self-supervised learning manner, SCL-SASR can easily super-resolve new images with arbitrary integer or non-integer scaling factors. Experiments on real-world images demonstrate the high flexibility and good applicability of SCL-SASR, which achieves better reconstruction performance than state-of-the-art self-supervised learning-based SISR methods as well as several external dataset-trained SISR models.

Progressive representation recalibration for lightweight super-resolution

Recently, the lightweight single-image super-resolution (SISR) task has received increasing attention due to the computational complexities and sizes of convolutional neural network (CNN)-based SISR models and the explosive demand in applications on resource-limited edge devices. Current algorithms reduce the number of layers and channels in CNNs to obtain lightweight models for this task. However, these algorithms may reduce the representation ability of the learned features due to information loss, inevitably leading to poor performance. In this work, we propose the progressive representation recalibration network (PRRN), a new lightweight SISR network to learn complete and representative feature representations. Specifically, a progressive representation recalibration block (PRRB) is developed to extract useful features from pixel and channel spaces in a two-stage approach. In the first stage, PRRB utilizes pixel and channel information to explore important feature regions. In the second stage, channel attention is further used to adjust the distribution of important feature channels. In addition, current channel attention mechanisms utilize nonlinear operations that may lead to information loss. In contrast, we design a shallow channel attention (SCA) mechanism that can learn the importance of each channel in a simpler yet more efficient way. Extensive experiments demonstrate the superiority of the proposed PRRN.

Single-Image Super Resolution of Remote Sensing Images with Real-World Degradation Modeling

Limited resolution is one of the most important factors hindering the application of remote sensing images (RSIs). Single-image super resolution (SISR) is a technique to improve the spatial resolution of digital images and has attracted the attention of many researchers. In recent years, with the advancement of deep learning (DL) frameworks, many DL-based SISR models have been proposed and achieved state-of-the-art performance; however, most SISR models for RSIs use the bicubic downsampler to construct low-resolution (LR) and high-resolution (HR) training pairs. Considering that the quality of the actual RSIs depends on a variety of factors, such as illumination, atmosphere, imaging sensor responses, and signal processing, training on “ideal” datasets results in a dramatic drop in model performance on real RSIs. To address this issue, we propose to build a more realistic training dataset by modeling the degradation with blur kernels and imaging noises. We also design a novel residual balanced attention network (RBAN) as a generator to estimate super-resolution results from the LR inputs. To encourage RBAN to generate more realistic textures, we apply a UNet-shape discriminator for adversarial training. Both referenced evaluations on synthetic data and non-referenced evaluations on actual images were carried out. Experimental results validate the effectiveness of the proposed framework, and our model exhibits state-of-the-art performance in quantitative evaluation and visual quality. We believe that the proposed framework can facilitate super-resolution techniques from research to practical applications in RSIs processing.

Online Learning for Reference-Based Super-Resolution

Online learning is a method for exploiting input data to update deep networks in the test stage to derive potential performance improvement. Existing online learning methods for single-image super-resolution (SISR) utilize an input low-resolution (LR) image for the online adaptation of deep networks. Unlike SISR approaches, reference-based super-resolution (RefSR) algorithms benefit from an additional high-resolution (HR) reference image containing plenty of useful features for enhancing the input LR image. Therefore, we introduce a new online learning algorithm, using several reference images, which is applicable to not only RefSR but also SISR networks. Experimental results show that our online learning method is seamlessly applicable to many existing RefSR and SISR models, and that improves performance. We further present the robustness of our method to non-bicubic degradation kernels with in-depth analyses.

Flexible Style Image Super-Resolution Using Conditional Objective

Recent studies have significantly enhanced the performance of single-image super-resolution (SR) using convolutional neural networks (CNNs). While there can be many high-resolution (HR) solutions for a given input, most existing CNN-based methods do not explore alternative solutions during the inference. A typical approach to obtaining alternative SR results is to train multiple SR models with different loss weightings and exploit the combination of these models. Instead of using multiple models, we present a more efficient method to train a single adjustable SR model on various combinations of losses by taking advantage of multi-task learning. Specifically, we optimize an SR model with a conditional objective during training, where the objective is a weighted sum of multiple perceptual losses at different feature levels. The weights vary according to given conditions, and the set of weights is defined as a style controller. Also, we present an architecture appropriate for this training scheme, which is the Residual-in-Residual Dense Block equipped with spatial feature transformation layers. At the inference phase, our trained model can generate locally different outputs conditioned on the style control map. Extensive experiments show that the proposed SR model produces various desirable reconstructions without artifacts and yields comparable quantitative performance to state-of-the-art SR methods. Code and trained models are available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/seungho-snu/FxSR</uri> .

TDPN: Texture and Detail-Preserving Network for Single Image Super-Resolution.

Single image super-resolution (SISR) using deep convolutional neural networks (CNNs) achieves the state-of-the-art performance. Most existing SISR models mainly focus on pursuing high peak signal-to-noise ratio (PSNR) and neglect textures and details. As a result, the recovered images are often perceptually unpleasant. To address this issue, in this paper, we propose a texture and detail-preserving network (TDPN), which focuses not only on local region feature recovery but also on preserving textures and details. Specifically, the high-resolution image is recovered from its corresponding low-resolution input in two branches. First, a multi-reception field based branch is designed to let the network fully learn local region features by adaptively selecting local region features in different reception fields. Then, a texture and detail-learning branch supervised by the textures and details decomposed from the ground-truth high resolution image is proposed to provide additional textures and details for the super-resolution process to improve the perceptual quality. Finally, we introduce a gradient loss into the SISR field and define a novel hybrid loss to strengthen boundary information recovery and to avoid overly smooth boundary in the final recovered high-resolution image caused by using only the MAE loss. More importantly, the proposed method is model-agnostic, which can be applied to most off-the-shelf SISR networks. The experimental results on public datasets demonstrate the superiority of our TDPN on most state-of-the-art SISR methods in PSNR, SSIM and perceptual quality. We will share our code on https://github.com/tocaiqing/TDPN.

Conditional Hyper-Network for Blind Super-Resolution With Multiple Degradations.

Although the single-image super-resolution (SISR) methods have achieved great success on the single degradation, they still suffer performance drop with multiple degrading effects in real scenarios. Recently, some blind and non-blind models for multiple degradations have been explored. However, these methods usually degrade significantly for distribution shifts between the training and test data. Towards this end, we propose a novel conditional hyper-network framework for super-resolution with multiple degradations (named CMDSR), which helps the SR framework learn how to adapt to changes in the degradation distribution of input. We extract degradation prior at the task-level with the proposed ConditionNet, which will be used to adapt the parameters of the basic SR network (BaseNet). Specifically, the ConditionNet of our framework first learns the degradation prior from a support set, which is composed of a series of degraded image patches from the same task. Then the adaptive BaseNet rapidly shifts its parameters according to the conditional features. Moreover, in order to better extract degradation prior, we propose a task contrastive loss to shorten the inner-task distance and enlarge the cross-task distance between task-level features. Without predefining degradation maps, our blind framework can conduct one single parameter update to yield considerable improvement in SR results. Extensive experiments demonstrate the effectiveness of CMDSR over various blind, and even several non-blind methods. The flexible BaseNet structure also reveals that CMDSR can be a general framework for a large series of SISR models. Our code is available at https://github.com/guanghaoyin/CMDSR.

Deep Stereoscopic Image Super-Resolution via Interaction Module

Deep learning-based methods have achieved remarkable performance in single image super-resolution. However, these methods cannot be effectively applied in stereoscopic image super-resolution without considering the characteristics of stereoscopic images. In this article, an interaction module-based stereoscopic image super-resolution network (IMSSRnet) is proposed to effectively utilize the correlation information in stereoscopic images. The key insight of the network lies with how to explore the complementary information of one view to help the reconstruction of another view. Thus, an interaction module is designed to acquire the enhanced features by utilizing complementary information between different views. Specifically, the interaction module is composed of a series of interaction units with a residual structure. In addition, the single image features of left and right views are obtained by a spatial feature extraction module, which can be realized by any existing single image super-resolution models. In order to obtain high-quality stereoscopic images, a gradient loss is introduced to preserve the texture details in a view, and a disparity loss is developed to constrain the disparity relationship between different views. Experimental results demonstrate that the proposed method achieves a promising performance and outperforms the state-of-the-art methods.