Domain Adaptation Research Articles

Health text simplification: An annotated corpus for digestive cancer education and novel strategies for reinforcement learning

Objective:The reading level of health educational materials significantly influences the understandability and accessibility of the information, particularly for minoritized populations. Many patient educational resources surpass widely accepted standards for reading level and complexity. There is a critical need for high-performing text simplification models for health information to enhance dissemination and literacy. This need is particularly acute in cancer education, where effective prevention and screening education can substantially reduce morbidity and mortality. Methods:We introduce Simplified Digestive Cancer (SimpleDC), a parallel corpus of cancer education materials tailored for health text simplification research, comprising educational content from the American Cancer Society, Centers for Disease Control and Prevention, and National Cancer Institute. The corpus includes 31 web pages with the corresponding manually simplified versions. It consists of 1183 annotated sentence pairs (361 train, 294 development, and 528 test). Utilizing SimpleDC and the existing Med-EASi corpus, we explore Large Language Model (LLM)-based simplification methods, including fine-tuning, reinforcement learning (RL), reinforcement learning with human feedback (RLHF), domain adaptation, and prompt-based approaches. Our experimentation encompasses Llama 2, Llama 3, and GPT-4. We introduce a novel RLHF reward function featuring a lightweight model adept at distinguishing between original and simplified texts when enables training on unlabeled data. Results:Fine-tuned Llama models demonstrated high performance across various metrics. Our RLHF reward function outperformed existing RL text simplification reward functions. The results underscore that RL/RLHF can achieve performance comparable to fine-tuning and improve the performance of fine-tuned models. Additionally, these methods effectively adapt out-of-domain text simplification models to a target domain. The best-performing RL-enhanced Llama models outperformed GPT-4 in both automatic metrics and manual evaluation by subject matter experts. Conclusion:The newly developed SimpleDC corpus will serve as a valuable asset to the research community, particularly in patient education simplification. The RL/RLHF methodologies presented herein enable effective training of simplification models on unlabeled text and the utilization of out-of-domain simplification corpora.

Towards Discriminative Class-Aware Domain Alignment via Coding Rate Reduction for Unsupervised Adversarial Domain Adaptation

Unsupervised domain adaptation (UDA) methods, based on adversarial learning, employ the means of implicit global and class-aware domain alignment to learn the symmetry between source and target domains and facilitate the transfer of knowledge from a labeled source domain to an unlabeled target domain. However, these methods still face misalignment and poor target generalization due to small inter-class domain discrepancy and large intra-class discrepancy of target features. To tackle these challenges, we introduce a novel adversarial learning-based UDA framework named Coding Rate Reduction Adversarial Domain Adaptation (CR2ADA) to better learn the symmetry between source and target domains. Integrating conditional domain adversarial networks with domain-specific batch normalization, CR2ADA learns robust domain-invariant features to implement global domain alignment. For discriminative class-aware domain alignment, we propose the global and local coding rate reduction methods in CR2ADA to maximize inter-class domain discrepancy and minimize intra-class discrepancy of target features. Additionally, CR2ADA combines minimum class confusion and mutual information to further regularize the diversity and discriminability of the learned features. The effectiveness of CR2ADA is demonstrated through experiments on four UDA datasets. The code can be obtained through email or GitHub.

Developmental Milestones for Children With Down Syndrome.

The American Academy of Pediatrics recommends that children and adolescents with Down syndrome receive anticipatory guidance regarding development and behavior. However, few tools provide specific guidance on developmental norms for children with Down syndrome. Our objective was to estimate age ranges at which children and adolescents with Down syndrome achieve developmental milestones to facilitate developmental screening by pediatric practitioners. We used standardized questionnaires to obtain information from clinicians and caregivers of children with Down syndrome who received care at the Boston Children's Hospital Down Syndrome Program between March 2018 and March 2023. Data included information from 2599 visits for 842 individuals with Down syndrome ages 2 months to 24 years. We used mixed-effects logistic regressions to predict the probability of achieving 25 specific developmental milestones with 15%, 30%, 45%, 60%, 75%, and 90% probability as a function of age. We further stratified results by individuals' sex. We present age norms for our study's population of people with Down syndrome for key milestones in academic, adaptive, language, and motor domains by calculating the ages at which milestone achievement was 75% probable. We then compare these norms to published norms for the general population. This study provides clinicians and families with age-based norms for achievement of key developmental milestones for children and adolescents with Down syndrome.

Joint weight optimization for partial domain adaptation via kernel statistical distance estimation

The goal of Partial Domain Adaptation (PDA) is to transfer a neural network from a source domain (joint source distribution) to a distinct target domain (joint target distribution), where the source label space subsumes the target label space. To address the PDA problem, existing works have proposed to learn the marginal source weights to match the weighted marginal source distribution to the marginal target distribution. However, this is sub-optimal, since the neural network’s target performance is concerned with the joint distribution disparity, not the marginal distribution disparity. In this paper, we propose a Joint Weight Optimization (JWO) approach that optimizes the joint source weights to match the weighted joint source distribution to the joint target distribution in the neural network’s feature space. To measure the joint distribution disparity, we exploit two statistical distances: the distribution-difference-based L2-distance and the distribution-ratio-based χ2-divergence. Since these two distances are unknown in practice, we propose a Kernel Statistical Distance Estimation (KSDE) method to estimate them from the weighted source data and the target data. Our KSDE method explicitly expresses the two estimated statistical distances as functions of the joint source weights. Therefore, we can optimize the joint weights to minimize the estimated distance functions and reduce the joint distribution disparity. Finally, we achieve the PDA goal by training the neural network on the weighted source data. Experiments on several popular datasets are conducted to demonstrate the effectiveness of our approach. Intro video and Pytorch code are available at https://github.com/sentaochen/Joint-Weight-Optimation. Interested readers can also visit https://github.com/sentaochen for more source codes of the related domain adaptation, multi-source domain adaptation, and domain generalization approaches.

A Domain Generation Diagnosis Framework for Unseen Conditions Based on Adaptive Feature Fusion and Augmentation

Emerging deep learning-based fault diagnosis methods have advanced in the current industrial scenarios of various working conditions. However, the prerequisite of obtaining target data in advance limits the application of these models to practical engineering scenarios. To address the challenge of fault diagnosis under unseen working conditions, a domain generation framework for unseen conditions fault diagnosis is proposed, which consists of an Adaptive Feature Fusion Domain Generation Network (AFFN) and a Mix-up Augmentation Method (MAM) for both the data and domain spaces. AFFN is utilized to fuse domain-invariant and domain-specific representations to improve the model’s generalization performance. MAM enhances the model’s exploration ability for unseen domain boundaries. The diagnostic framework with AFFN and MAM can effectively learn more discriminative features from multiple source domains to perform different generalization tasks for unseen working loads and machines. The feasibility of the proposed unseen conditions diagnostic framework is validated on the SDUST and PU datasets and achieved peak diagnostic accuracies of 94.15% and 93.27%, respectively.

Subfield-level crop yield mapping without ground truth data: A scale transfer framework

Ongoing advances in satellite remote sensing data and machine learning methods have enabled crop yield estimation at various spatial and temporal resolutions. While yield mapping at broader scales (e.g., state or county level) has become common, mapping at finer scales (e.g., field or subfield) has been limited by the lack of ground truth data for model training and evaluation. Here we present a scale transfer framework, named Quantile loss Domain Adversarial Neural Networks (QDANN), that leverages knowledge from county-level datasets to map crop yields at the subfield level. Based on the strategy of unsupervised domain adaptation, QDANN is trained on labeled county-level data and unlabeled subfield-level data, with no requirement for yield information at the subfield level. We evaluate the proposed method applied to Landsat imagery and Gridmet weather data for maize, soybean, and winter wheat fields in the United States, using as reference data yield monitor records from roughly one million field-year observations. The model is compared with several process-based and machine learning-based benchmark approaches that train on simulated yield records or county-level data. QDANN-estimated yields achieved an R2 score (RMSE) of 48 % (2.29 t/ha), 32 % (0.85 t/ha), and 39 % (1.40 t/ha) for maize, soybean, and winter wheat in comparison with the ground-based yield measures, respectively. These performances are higher than benchmark approaches and are nearly as good as models trained on field-level data. When aggregated to the county level, the improvement achieved by QDANN is more pronounced and the R2 scores (RMSE) improved to 78 % (0.98 t/ha), 62 % (0.37 t/ha), and 53 % (1.00 t/ha) for maize, soybean, and winter wheat, respectively. This study demonstrates that the proposed scale transfer framework can serve as a reliable approach for yield mapping at the subfield level when there is no access to fine-scale yield information. Based on the QDANN model, we have generated and made publicly available 30-m annual yield maps for major crop-producing states in the U.S. since 2008.

Taking it further: Leveraging pseudo-labels for field delineation across label-scarce smallholder regions

Satellite-based field delineation has entered a quasi-operational stage due to recent advances in machine learning for computer vision. Transfer learning allows for the resource-efficient transfer of pre-trained field delineation models across heterogeneous geographies. However, the scarcity of labeled data for complex and dynamic smallholder landscapes remains a major bottleneck. The key innovation of this study is to overcome this challenge by using pre-trained models to generate sparse (i.e., not fully annotated) field delineation pseudo-labels for fine-tuning models across geographies and sensor characteristics. We build on a FracTAL ResUNet trained for crop field delineation in India (median field size of 0.24 ha) based on multi-spectral imagery at 1.5 m spatial resolution. We use this model to generate pseudo-labels for the use in Northern Mozambique (median field size of 0.06 ha) based on sub-meter resolution true-color satellite imagery. We designed multiple pseudo-label selection strategies based on field-level probability scores and compared the quantities, area properties, seasonal distribution, and spatial agreement of the pseudo-labels against human-annotated training labels (n = 1,512). We then used the human-annotated labels and the pseudo-labels for model fine-tuning and compared predictions against human field annotations (n = 2,199). We evaluated performance with regards to object-level spatial agreement and site-level field size estimation. Our results indicate i) a good baseline performance of the pre-trained model in both field delineation (mean intersection over union (mIoU) of 0.634) and field size estimation (mean root mean squared error (mRMSE) of 0.071 ha), and ii) the added value of regional fine-tuning with performance improvements in nearly all experiments (mIoU increases of up to 0.060, mRMSE decreases of up to 0.034 ha). Moreover, we found iii) substantial performance increases when using only pseudo-labels (up to 77 % of the mIoU increases and 68 % of the mRMSE decreases obtained by human-annotated labels), and iv) additional performance increases (mIoU+0.008, mRMSE: −0.003 ha) when complementing human annotations with pseudo-labels. Pseudo-labels are architecture-agnostic, can be efficiently generated at scale, and thus facilitate domain adaptation in label-scarce settings. The workflow presented here is a stepping stone for overcoming the persisting challenges in mapping heterogeneous smallholder agriculture.

Unsupervised domain adaptation bearing fault diagnosis method based on joint feature alignment

Unsupervised domain adaptation bearing fault diagnosis method based on joint feature alignment

Label distribution learning for compound facial expression recognition in‐the‐wild: A comparative study

AbstractHuman emotional states encompass both basic and compound facial expressions. However, current works primarily focus on basic expressions, consequently neglecting the broad spectrum of human emotions encountered in practical scenarios. Compound facial expressions involve the simultaneous manifestation of multiple emotions on an individual's face. This phenomenon reflects the complexity and richness of human states, where facial features dynamically convey a combination of feelings. This study embarks on a pioneering exploration of Compound Facial Expression Recognition (CFER), with a distinctive emphasis on leveraging the Label Distribution Learning (LDL) paradigm. This strategic application of LDL aims to address the ambiguity and complexity inherent in compound expressions, marking a significant departure from the dominant Single Label Learning (SLL) and Multi‐Label Learning (MLL) paradigms. Within this framework, we rigorously investigate the potential of LDL for a critical challenge in Facial Expression Recognition (FER): recognizing compound facial expressions in uncontrolled environments. We utilize the recently introduced RAF‐CE dataset, meticulously designed for compound expression assessment. By conducting a comprehensive comparative analysis pitting LDL against conventional SLL and MLL approaches on RAF‐CE, we aim to definitively establish LDL's superiority in handling this complex task. Furthermore, we assess the generalizability of LDL models trained on RAF‐CE by evaluating their performance on the EmotioNet and RAF‐DB Compound datasets. This demonstrates their effectiveness without domain adaptation. To solidify these findings, we conduct a comprehensive comparative analysis of 12 cutting‐edge LDL algorithms on RAF‐CE, S‐BU3DFE, and S‐JAFFE datasets, providing valuable insights into the most effective LDL techniques for FER in‐the‐wild.

Advancing Nighttime Object Detection through Image Enhancement and Domain Adaptation

Due to the lack of annotations for nighttime low-light images, object detection in low-light images has always been a challenging problem. Achieving high-precision results at night is also an issue. Additionally, we aim to use a single nighttime dataset to complete the knowledge distillation task while improving the detection accuracy of object detection models under nighttime low-light conditions and reducing the computational cost of the model, especially for small targets and objects contaminated by special nighttime lighting. This paper proposes a Nighttime Unsupervised Domain Adaptation Network (NUDN) based on knowledge distillation to address these issues. To improve the detection accuracy of nighttime images, high-confidence bounding box predictions from the teacher and region proposals from the student are first fused, allowing the teacher to perform better in subsequent training, thus generating a combination of high-confidence and low-confidence pseudo-labels. This combination of feature information is used to guide model training, enabling the model to extract feature information similar to that of source images in nighttime low-light images. Nighttime images and pseudo-labels undergo random size transformations before being used as input for the student, enhancing the model’s generalization across different scales. To address the scarcity of nighttime datasets, we propose a nighttime-specific augmentation pipeline called LightImg. This pipeline enhances nighttime features, transforming them into daytime features and reducing issues such as backlighting, uneven illumination, and dim nighttime light, enabling cross-domain research using existing nighttime datasets. Our experimental results show that NUDN can significantly improve nighttime low-light object detection accuracy on the SHIFT and ExDark datasets. We conduct extensive experiments and ablation studies to demonstrate the effectiveness and efficiency of our work.

Harmony in diversity: Content cleansing change detection framework for very-high-resolution remote-sensing images

Change detection, as a crucial task in the field of Earth observation, aims to identify changed pixels between multi-temporal remote-sensing images captured at the same geographical area. However, in practical applications, there are challenges of pseudo changes arising from diverse imaging conditions and different remote-sensing platforms. Existing methods either overlook the different imaging styles between bi-temporal images, or transfer the bi-temporal styles via domain adaptation that may lose ground details. To address these problems, we introduce the disentangled representation learning that mitigates differences of imaging styles while preserving content details to develop a change detection framework, named Content Cleansing Network (CCNet). Specifically, CCNet embeds each input image into two distinct subspaces: a shared content space and a private style space. The separation of style space aims to mitigate the discrepant style due to different imaging condition, while the extracted content space reflects semantic features that is essential for change detection. Then, a multi-resolution parallel structure constructs the content space encoder, facilitating robust feature extraction of semantic information and spatial details. The cleansed content features enable accurate detection of changes in the land surface. Additionally, a lightweight decoder for image restoration enhances the independence and interpretability of the disentangled spaces. To verify the proposed method, CCNet is applied to five public datasets and a multi-temporal dataset collected in this study. Comparative experiments against eleven advanced methods demonstrate the effectiveness and superiority of CCNet. The experimental results show that our method robustly addresses the issues related to both temporal and platform variations, making it a promising method for change detection in complex conditions and supporting downstream applications.

RT-SNDETR: real-time supernova detection via end-to-end image transformers

ABSTRACT In large-scale astronomical surveys, traditional supernova detection pipelines rely on complex and relatively inefficient image differencing techniques. This paper proposes an end-to-end deep-learning supernova detection network, the Real-Time SuperNova DEtection TRansformer (RT-SNDETR). This network partially replaces traditional pipelines by integrating image differencing, source detection, and Real-bogus classification, achieving a speed 51.49 times that of the fastest image differencing method, SFFT. Additionally, it remains competitive with methods like YOLO v8, offering a well-balanced trade-off between speed and accuracy. Experimental results highlight RT-SNDETR’s superior performance, with an average precision(AP) of 96.30 per cent on synthetic samples and 76.60 per cent on real supernova samples. It significantly outperforms other detection networks, including RT-DETR (+5.6 per cent AP on synthetic/+5.1 per cent AP on real samples) and Cascade R-CNN (+8.9 per cent AP on synthetic/ +28.6 per cent AP on real samples). The incorporation of CycleGAN-based data generation methods plays a significant role in enhancing RT-SNDETR’s performance. These methods simulate realistic PSF variations, enabling the object detection network to learn more robust features and improving its generalization to real supernovae data. Additionally, by integrating unsupervised domain adaptation techniques, RT-SNDETR achieves an AP of 81.70 per cent on real SDSS supernova survey samples. This study demonstrates RT-SNDETR’s potential to significantly enhance both the speed and accuracy of supernova detection, making it a highly effective solution for large-scale astronomical surveys.

USDAP: universal source-free domain adaptation based on prompt learning

USDAP: universal source-free domain adaptation based on prompt learning

Cross-pollination of knowledge for object detection in domain adaptation for industrial automation

AbstractArtificial Intelligence is revolutionizing industries by enhancing efficiency through real-time Object Detection (OD) applications. Utilizing advanced computer vision techniques, OD systems automate processes, analyze complex visual data, and facilitate data-driven decisions, thus increasing productivity. Domain Adaptation for OD has recently gained prominence for its ability to recognize target objects without annotations. Innovative approaches that merge traditional cross-disciplinary domain modeling with cutting-edge deep learning have become essential in addressing complex AI challenges in real-time scenarios. Unlike traditional methods, this study proposes a novel, effective Cross-Pollination of Knowledge (CPK) strategy for domain adaptation inspired by botanical processes. The CPK approach involves merging target samples with source samples at the input stage. By incorporating a random and unique selection of a few target samples, the merging process enhances object detection results efficiently in domain adaptation, supporting detectors in aligning and generalizing features with the source domain. Additionally, this work presents the new Planeat digit recognition dataset, which includes 231 images. To ensure robust comparison, we employ a self-supervised Domain Adaptation (UDA) method that simultaneously trains target and source domains using unsupervised techniques. UDA method leverages target data to identify high-confidence regions, which are then cropped and augmented, adapting UDA for effective OD. The proposed CPK approach significantly outperforms existing UDA techniques, improving mean Average Precision (mAP) by 10.9% through rigorous testing on five diverse datasets across different conditions- cross-weather, cross-camera, and synthetic-to-real. Our code is publicly available https://github.com/anwaar0/CPK-Object-Detection

Crowd Counting Using Meta-Test-Time Adaptation.

Machine learning algorithms are commonly used for quickly and efficiently counting people from a crowd. Test-time adaptation methods for crowd counting adjust model parameters and employ additional data augmentation to better adapt the model to the specific conditions encountered during testing. The majority of current studies concentrate on unsupervised domain adaptation. These approaches commonly perform hundreds of epochs of training iterations, requiring a sizable number of unannotated data of every new target domain apart from annotated data of the source domain. Unlike these methods, we propose a meta-test-time adaptive crowd counting approach called CrowdTTA, which integrates the concept of test-time adaptation into the meta-learning framework and makes it easier for the counting model to adapt to the unknown test distributions. To facilitate the reliable supervision signal at the pixel level, we introduce uncertainty by inserting the dropout layer into the counting model. The uncertainty is then used to generate valuable pseudo labels, serving as effective supervisory signals for adapting the model. In the context of meta-learning, one image can be regarded as one task for crowd counting. In each iteration, our approach is a dual-level optimization process. In the inner update, we employ a self-supervised consistency loss function to optimize the model so as to simulate the parameters update process that occurs during the test phase. In the outer update, we authentically update the parameters based on the image with ground truth, improving the model's performance and making the pseudo labels more accurate in the next iteration. At test time, the input image is used for adapting the model before testing the image. In comparison to various supervised learning and domain adaptation methods, our results via extensive experiments on diverse datasets showcase the general adaptive capability of our approach across datasets with varying crowd densities and scales.

Beyond RNAi: How the Dicer protein modulates the antiviral innate immune response in mammalian cells: Mammalian Dicer could regulate the innate immune response in an RNAi-independent manner as a result of losing long dsRNA processive activity.

While Dicer plays an important antiviral role through the RNAi pathway in plants and invertebrates, its contribution to antiviral immunity in vertebrates and more specifically mammals is more controversial. The apparent limited RNAi activity in mammalian cells has been attributed to the reduced long dsRNA processive activity of mammalian Dicer, as well as a functional incompatibility between the RNAi and IFN pathways. Why Dicer has lost this antiviral activity in the profit of the IFN pathway is still unclear. We propose that the primary direct antiviral activity of Dicer has been functionally replaced by other sensors in the IFN pathway, leading to its specialization toward microRNA maturation. As a result, Dicer can regulate the innate immune response and prevent basal activation of the IFN pathway in mammals. Here, we discuss this hypothesis, highlighting how the adaptation of the helicase domain of mammalian Dicer may be key to this process.

Fault diagnosis method for bearings under variable working conditions based on transfer relation network

Abstract In order to achieve robust fault diagnosis under varying conditions with limited labeled data, this study combines metric-based meta-learning with feature-based domain adaptation. It introduces a new approach for variable-condition bearing fault diagnosis using transfer relation networks. To enhance the network’s ability to generalize across different domains, the paper integrates local maximum mean discrepancy (LMMD) into the relation network architecture. LMMD aligns the data distributions of various classes between the source and target domains, effectively addressing distributional differences and improving model generalization. To accurately and swiftly extract meaningful fault features, the study proposes a lightweight feature extraction module based on Shuffle Attention (SA). This module employs depth-wise separable convolutions for efficiency and integrates SA after each convolutional layer to bolster feature representation. Finally, experiments on two bearing datasets under varying conditions validate the efficacy and superiority of the proposed model over alternative methods.

Multi-source domain adaptation network for partial discharge severity assessment in gas-insulated switchgear

Abstract Gas-insulated switchgear (GIS) partial discharge (PD) severity assessment is critical for ensuring the reliable operation of GIS systems. However, existing assessment methods often overlook the long-term dependencies of historical data and fail to adequately address challenges related to limited on-site samples and variations in sample distribution. To overcome these challenges, we propose a novel multi-source domain adaptation network (MSDAN) specifically designed for on-site GIS PD severity assessment, which is the first model developed considering distribution differences in GIS PD severity assessment for different defect types. Our approach begins with the development of a feature extractor that captures both discernible PD features and long-term dependencies. We then introduce a multi-source domain adaptation strategy to mitigate distribution disparities across PD severity samples from different defect types, ensuring effective domain alignment. Additionally, we incorporate an adaptive weighted classification mechanism to accurately assess PD severity by considering the varying contributions of different defect types to the target domain task. Experimental results demonstrate that MSDAN achieves a remarkable accuracy of 95.38% in on-site GIS PD severity assessment, outperforming other benchmark models in both accuracy and robustness. This highlights the potential of MSDAN as a robust solution for real-world GIS PD severity assessment.

A global reweighting approach for cross-domain semantic segmentation

Unsupervised domain adaptation semantic segmentation attracts much research attention due to the expensive pixel-level annotation cost. Since the adaptation difficulty of samples is different, the weight of samples should be set independently, which is called reweighting. However, existing reweighting methods only calculate local reweighting information from predicted results or context information in batch images of two domains, which may lead to over-alignment or under-alignment problems. To handle this issue, we propose a global reweighting approach. Specifically, we first define the target centroid distance, which describes the distance between the source batch data and the target centroid. Then, we employ a Fréchet Inception Distance metric to evaluate the domain divergence and embed it into the target centroid distance. Finally, a global reweighting strategy is proposed to enhance the knowledge transferability in the source domain supervision. Extensive experiments demonstrate that our approach achieves competitive performance and helps to improve performance in other methods.

Domain adaptation hyperspectral image fusion based on spatial-spectral domain separation

ABSTRACT Hyperspectral datasets captured by airborne are one of the important data sources in the field of hyperspectral image fusion. However, the limited number of data samples often hinders the optimal performance of deep learning-based methods on these datasets. Domain adaptation is expected to alleviate the issue of data scarcity. However, there are two issues that need to be addressed in applying domain adaptation to hyperspectral image fusion, which are the heterogeneity and spatial degradation diversity of hyperspectral data. In this paper, we propose a domain adaptation hyperspectral image fusion network based on spatial-spectral domain separation. The model, constructed with three-dimensional convolutional layers, effectively addresses the challenge of hyperspectral data heterogeneity. Compared with 2D convolution, 3D convolution is able to consider both spatial and spectral dimensions and extract spatial-spectral features efficiently. Based on this, we design a domain-separation architecture to extract domain-invariant and domain-private features in the spatial-spectral domain. The architecture achieves domain adaptation by extracting domain-invariant features from both datasets to obtain the prior knowledge in the source dataset and apply it to the target dataset. Additionally, the spatial degradation of different datasets varies due to distinct shooting conditions during the acquisition. Therefore, the fusion algorithms should be robust to various degradation scenarios. To tackle this challenge, we design the Observation Module to predict degradation information and the Degradation Information Modulation Module to apply it to the input, thereby enhancing the network’s robustness. Experiments on various datasets demonstrate that our method is qualitatively and quantitatively superior to state-of-the-art methods.