Target Domain Dataset Research Articles

Arc bubble edge detection method based on deep transfer learning in underwater wet welding.

The stability of arc bubble is a crucial indicator of underwater wet welding process. However, limited research exists on detecting arc bubble edges in such environments, and traditional algorithms often produce blurry and discontinuous results. To address these challenges, we propose a novel arc bubble edge detection method based on deep transfer learning for processing underwater wet welding images. The proposed method integrates two training stages: pre-training and fine-tuning. In the pre-training stage, a large source domain dataset is used to train VGG16 as a feature extractor. In the fine-tuning stage, we introduce the Attention-Scale-Semantics (ASS) model, which consists of a Convolutional Block Attention Module (CBAM), a Scale Fusion Module (SCM) and a Semantic Fusion Module (SEM). The ASS model is further trained on a small target domain dataset specific to underwater wet welding to fine-tune the model parameters. The CBAM can adaptively weight the feature maps, focusing on more crucial features to better capture edge information. The SCM training method maximizes feature utilization and simplifies training by combining multi-scale features. Additionally, the skip structure of SEM effectively mitigates semantic loss in the high-level network, enhancing the accuracy of edge detection. On the BSDS500 dataset and a self-constructed underwater wet welding dataset, the ASS model was evaluated against conventional edge detection models-Richer Convolutional Features (RCF), Fully Convolutional Network (FCN), and UNet-as well as state-of-the-art models LDC and TEED. In terms of Mean Absolute Error (MAE), accuracy, and other evaluation metrics, the ASS model consistently outperforms these models, demonstrating edge detection capabilities that are both effective and stable in detecting arc bubble edges in underwater wet welding images.

Multisource Modeling Method for Petroleum Production Prediction Models: A Case Study of CO2-Flooding Performance

Summary Accurate prediction of oil production is crucial for formulating oilfield development strategies. With the rapid development of artificial intelligence, research on utilizing deep learning to construct oil production prediction models has been growing, which has partially compensated for the low computational efficiency of numerical simulators. Although the well-trained source domain model maintains high prediction accuracy on target blocks with similar production conditions, the prediction accuracy of the model declines in scenarios where substantial disparities exist between the production conditions of the target block and the source domain. This discrepancy makes the prediction results unreliable and causes a domain shift issue. We propose a multisource model fine-tuning approach, which leverages a limited amount of target domain data to fine-tune the existing source domain model, enabling it to rapidly converge in the target domain while maintaining superior prediction performance. Based on a heterogeneous low-permeability CO2-flooding reservoir development model, we established a series of source domain data sets, encompassing numerous types of well patterns and permeability fields, and specifically prepared various target domain data sets to verify the effectiveness of the model fine-tuning. Experimental outcomes demonstrate that our proposed model fine-tuning approach facilitates the rapid convergence of the existing model on target domain data. Following testing, the fine-tuned model, which attained a prediction accuracy exceeding 97% in the target domain, significantly improved upon the accuracy compared with the unfine-tuned model. The time required is significantly lower than retraining a new model, and it significantly reduces the need for data in the target domain. This provides support for the rapid generation of new prediction models using existing source domain models and limited target domain data.

MURDA: Multisource Unsupervised Raman Spectroscopy Domain Adaptation Model with Reconstructed Target Domains for Medical Diagnosis Assistance.

Artificial intelligence combined with Raman spectroscopy for disease diagnosis is on the rise. However, these methods require a large amount of annotated spectral data for modeling to achieve high diagnostic accuracy. Annotating labels consumes significant medical resources and time. To reduce dependence on labeled medical data resources, we propose a method called Multisource Unsupervised Raman Spectroscopy Domain Adaptation Model with Reconstructed Target Domains (MURDA). It transfers knowledge learned from source domain data sets of different diseases to an unlabeled target domain data set. Compared to knowledge transfer from a single source domain, knowledge from multiple disease source domains provides more generalized knowledge. Considering the diversity of autoimmune diseases and the limited sample size, we apply MURDA to assist in the medical diagnosis of autoimmune diseases. Additionally, we propose a Double-Branch Multiscale Convolutional Self-Attention (DMCS) feature extractor that is more suitable for spectral data feature extraction. On three sets of serum Raman spectroscopy data sets for autoimmune diseases, the multisource domain adaptation diagnostic accuracy of MURDA was superior to traditional single source and multisource models, with accuracy rates of 73.6%, 83.4%, and 82.9%, respectively. Compared with pure source tasks without domain adaptation, it improved by 15.1%, 36%, and 21.6%, respectively. This demonstrates the effectiveness of Raman spectroscopy combined with MURDA in diagnosing autoimmune diseases. We investigated the important decision dependency peaks in migration tasks, providing assistance for future research on artificial intelligence combined with Raman spectroscopy for diagnosing autoimmune diseases. Furthermore, to validate the effectiveness and generalization performance of MURDA, we conducted experiments on the publicly available RRUFF data set, exploring the application of multisource unsupervised domain adaptation in more Raman spectroscopy scenarios.

General framework for unsteady aerodynamic prediction of airfoils based on deep transfer learning

Analyzing the unsteady aerodynamic performance of airfoils under dynamic stall using computational fluid dynamics (CFD) is computationally intensive. Although deep learning models can quickly predict aerodynamic parameters, their generalization capability on a small-scale dataset is often poor. This paper presents a novel deep transfer learning (TL) framework that combines model-based and instance-based transfer methods, termed synergistic instance-model TL. This framework facilitates rapid predictions of unsteady aerodynamic performance for various airfoils and pitch oscillations from the small-scale dataset. The framework integrates the accelerated training speed of model-based methods with the dynamic dataset expansion benefits of instance-based approaches. Initially, a pre-trained Wasserstein-deep convolutional generative adversarial network (W-DCGAN) is developed, combining a convolutional neural network with a generative adversarial network to predict aerodynamic hysteresis loops for the SC1095 airfoil in the source domain. The framework then fine-tunes the pre-trained model and incorporates weighted source domain dataset into the small-scale target domain dataset, producing the transferred model W-DCGAN-TL. This approach significantly reduces prediction inaccuracies compared to model-based and non-TL methods when applied to the small-scale dataset. The framework's flexibility allows the use of pre-trained models and datasets from related aerodynamic problems to address issues with insufficient data. Consequently, it is expected to reduce the dependency on extensive datasets, enhance design efficiency, and minimize resource requirements.

An optimized heterogeneous multi-access edge computing framework based on transfer learning and artificial internet of things

In most practical applications, the feature space of the training datasets and the target domain datasets are inconsistent, or the data distribution between them is inconsistent, which leads to the problem of data starvation and makes it difficult for terminal devices to obtain high accurate results. Aiming at the problems of limited terminal device resources, low accuracy of data processing results, and unsatisfactory processing speed, a Heterogeneous Multi-access Edge Computing (MEC) Framework based on Transfer Learning (TL) is proposed, abbreviated as HMECF-TL. This framework adopts a cloud-edge-end three-layer architecture. It uses model transfer to optimize the Convolutional Neural Networks (CNN) model at each layer to achieve the goal of improving data processing speed and accuracy. Furthermore, a multi-agent Deep Reinforcement Learning Algorithm having Attention Mechanism (DRLAAM) is designed to further increase the timeliness performance of computation-intensive applications. The performance of HMECSF-TL framework is verified by simulation experiments, which not only reduces the delay by more than 24.66 %, but also improves the accuracy by more than 8.34 %. The framework not only increase the computing capacity to solve the shortage of terminal device resources, but also improve the quality of data processing to solve the problem of data starvation.

In the Wild Video Violence Detection: An Unsupervised Domain Adaptation Approach

This work addresses the challenge of video violence detection in data-scarce scenarios, focusing on bridging the domain gap that often hinders the performance of deep learning models when applied to unseen domains. We present a novel unsupervised domain adaptation (UDA) scheme designed to effectively mitigate this gap by combining supervised learning in the train (source) domain with unlabeled test (target) data. We employ single-image classification and multiple instance learning (MIL) to select frames with the highest classification scores, and, upon this, we exploit UDA techniques to adapt the model to unlabeled target domains. We perform an extensive experimental evaluation, using general-context data as the source domain and target domain datasets collected in specific environments, such as violent/non-violent actions in hockey matches and public transport. The results demonstrate that our UDA pipeline substantially enhances model performances, improving their generalization capabilities in novel scenarios without requiring additional labeled data.

MDTGAN: Multi domain generative adversarial transfer learning network for traffic data imputation

Accurate, real-time, and efficient traffic data, crucial for intelligent transportation systems, which is often disrupted in the real world due to the influence of weather, interference, and equipment failures. Generative Adversarial Networks (GANs), have achieved significant results in image restoration, providing insights for traffic data imputation. Recent research has focused on developing more effective and reasonable model by learning transferable common knowledge from different cities or different scenarios, which is also an excellent idea for improving the data imputation performance. In light of, we contrive to design a GANs based transferable traffic data imputation model, namely Multi Domain Generative Adversarial Transfer Learning Network (MDTGAN). Firstly, the model consists of two stages. In pre-training stage, the model utilizes the source domain datasets (Similar cities road network datasets) to update parameters and transferred them. Then, the model parameters are optimized using the target domain dataset (Target city road network dataset) in fine-tuning stage to avoid the issue of insufficient samples caused by data missing. Secondly, to adapt the model to different dataset topologies, MDTGAN models the node-level data by taking node time series as input, enabling the model to autonomously learn the prevalent spatiotemporal correlations inherent in nodes. Additionally, we introduce a domain discriminator module that guides the spatial encoder in learning domain-invariant features of network node spatial information, enhancing the model generalization ability. The experiments conducted on three publicly available datasets, in which one dataset is regarded as the target dataset, while the others serve as source datasets. The experimental results demonstrate that the MDTGAN model consistently outperforms the baseline models. Specifically, the MDTGAN model can transfer valuable knowledge from the source domain datasets to improve data imputation performance on the target domain dataset, providing practical significance for cities lacking historical traffic data.

Continuous unsupervised domain adaptation using stabilized representations and experience replay

We introduce an algorithm for tackling the problem of unsupervised domain adaptation (UDA) in continual learning (CL) scenarios. The primary objective is to maintain model generalization under domain shift when new domains with solely unannotated data arrive continually. Under this setting, While there are many existing UDA algorithms, we have access to only the current target domain dataset simultaneously and need to update the model to learn a task with unannotated data without forgetting the tasks learnt before. Our contribution is to address this learning setting which faces challenges seen in both UDA and CL. The core assumption in the learning setting is that the tasks share the same classes which enables benefiting from knowledge transfer. Our solution is based on stabilizing the learned internal distribution after learning the initial task to enhances the model generalization on new domains. The internal distribution is modeled by network responses in the eventual model hidden layer. We model this internal distribution using a Gaussian mixture model (GMM ) and update the model by matching the internally learned distribution of new domains to the estimated GMM as a surrogate for the past learned internal distribution. Additionally, we leverage experience replay to overcome the problem of catastrophic forgetting, where the model loses previously acquired knowledge when learning new tasks. We offer theoretical analysis to explain why our algorithm would work. We also offer comparative and analytic experiments to demonstrate that our method is effective. Our implementation code is available at: https://github.com/rostami-m/LDACID.

Stereo sample generation‐based domain generalization network for stereo matching

AbstractRecently, deep learning‐based stereo matching has achieved great success. However, models trained on the source domain dataset encounter substantial performance degradation when directly tested on an unseen target domain dataset because of neglecting the generalization to out‐of‐distribution (OOD) stereo samples. This paper proposes a stereo sample generation‐based domain generalization network (SGDG‐Net) for stereo matching. Specifically, to expand the distribution span of training samples, OOD stereo samples are generated to assist training. To effectively generate OOD left samples, a style transfer‐based generation mechanism is proposed to transmit perturbations to the source left samples. In addition, to generate the OOD right samples, a disparity‐assisted generation strategy is proposed by using disparity map labels as auxiliary information. Experimental results demonstrate that the proposed SGDG‐Net produces remarkable results on four benchmark datasets.

Cross-Scene Hyperspectral Image Classification Based on Graph Alignment and Distribution Alignment

A domain alignment-based hyperspectral image (HSI) classification method was designed to address the heterogeneity in resolution and band between the source domain and target domain datasets of cross-scene hyperspectral images, as well as the resulting reduction in common features. Firstly, after preliminary feature extraction, perform two domain alignment operations: image alignment and distribution alignment. Image alignment aims to align hyperspectral images of different bands or time points, ensuring that they are within the same spatial reference framework. Distribution alignment adjusts the distribution of features of samples of different categories in the feature space to reduce the distribution differences of the same type of features between two domains. Secondly, adjust the consistency of the two alignment methods to ensure that the features obtained through different alignment methods exhibit consistency in the feature space, thereby improving the comparability and reliability of the features. In addition, this method considers multiple losses in the model from different perspectives and makes comprehensive adjustments through a unified optimization process to more comprehensively capture and utilize the correlation information between data. Experimental results on Houston 2013 and Houston 2018 datasets can improve the hyperspectral prediction performance between datasets with different resolutions and bands, effectively solving the problems of high cost and limited training samples in HSI labeling and significantly improving cross-scene HSI classification performance.

SFDA-CD: A Source-Free Unsupervised Domain Adaptation for VHR Image Change Detection

Deep models may have disappointing performance in real applications due to the domain shifts in data distributions between the source and target domain. Although a few unsupervised domain adaptation methods have been proposed to make the pre-train models effective on target domain datasets, constraints like data privacy, security, and transmission limits restrict access to VHR remote sensing images, making existing unsupervised domain adaptation methods almost ineffective in specific change detection areas. Therefore, we propose a source-free unsupervised domain adaptation change detection structure to complete specific change detection tasks, using only the pre-trained source model and unlabelled target data. The GAN-based source generation component is designed to generate synthetic source data, which, to some extent, reflects the distribution of the source domain. Moreover, these data can be utilised in model knowledge transfer. The model adaptation component facilitates knowledge transfer between models by minimising the differences between deep features, using AAM (Attention Adaptation Module) to extract the difference between high-level features, meanwhile we proposed ISM (Intra-domain Self-supervised Module) to train target model in a self-supervised strategy in order to improve the knowledge adaptation. Our SFDA-CD framework demonstrates superior accuracy over existing unsupervised domain adaptation change detection methods, which has 0.6% cIoU and 1.5% F1 score up in cross-regional tasks and 1.4% cIoU and 1.9% F1 score up in cross-scenario tasks, proving that it can effectively reduce the domain shift between the source and target domains even without access to source data. Additionally, it can facilitate knowledge transfer from the source model to the target model.

Parameter-based transfer learning for severity classification of atopic dermatitis using hyperspectral imaging.

Because atopic dermatitis (AD) is a chronic inflammatory skin condition that causes structural changes, there is a growing need for noninvasive research methods to evaluate this condition. Hyperspectral imaging (HSI) captures skin structure features by exploiting light wavelength variations in penetration depth. In this study, parameter-based transfer learning was deployed to classify the severity of AD using HSI. Therefore, we aimed to obtain an optimal combination of classification results from the four models after constructing different source- and target-domain datasets. We designated psoriasis, skin cancer, eczema, and AD datasets as the source datasets, and the set of images acquired via hyperspectral camera as the target dataset for wavelength-specific AD classification. We compared the severity classification performances of 96 combinations of sources, models, and targets. The highest classification performance of 83% was achieved when ResNet50 was trained on the augmented psoriasis dataset as the source, with the resulting parameters used to train the model on the target Near-infrared radiation (NIR) dataset. The second highest classification accuracy of 81% was achieved when ResNet50 was trained on the unaugmented psoriasis dataset as the source, with the resulting parameters used to train the model on the target R dataset. ResNet50 demonstrated potential as a generalized model for both the source and target data, also confirming that the psoriasis dataset is an effective training resource. The present study not only demonstrates the feasibility of the severity classification of AD based on hyperspectral images, but also showcases combinations and research scalability for domain exploration.

A hybrid neural network model based on transfer learning for Arabic sentiment analysis of customer satisfaction

AbstractSentiment analysis, a method used to classify textual content into positive, negative, or neutral sentiments, is commonly applied to data from social media platforms. Arabic, an official language of the United Nations, presents unique challenges for sentiment analysis due to its complex morphology and dialectal diversity. Compared to English, research on Arabic sentiment analysis is relatively scarce. Transfer learning, which applies the knowledge learned from one domain to another, can address the limitations of training time and computational resources. However, the development of transfer learning for Arabic sentiment analysis is still underdeveloped. In this study, we develop a new hybrid model, RNN‐BiLSTM, which merges recurrent neural networks (RNN) and bidirectional long short‐term memory (BiLSTM) networks. We used Arabic bidirectional encoder representations from transformers (AraBERT), a state‐of‐the‐art Arabic language pre‐trained transformer‐based model, to generate word‐embedding vectors. The RNN‐BiLSTM model integrates the strengths of RNN and BiLSTM, including the ability to learn sequential dependencies and bidirectional context. We trained the RNN‐BiLSTM model on the source domain, specifically the Arabic reviews dataset (ARD). The RNN‐BiLSTM model outperforms the RNN and BiLSTM models with default parameters, achieving an accuracy of 95.75%. We further applied transfer learning to the RNN‐BiLSTM model by fine‐tuning its parameters using random search. We compared the performance of the fine‐tuned RNN‐BiLSTM model with the RNN and BiLSTM models on two target domain datasets: ASTD and Aracust. The results showed that the fine‐tuned RNN‐BiLSTM model is more effective for transfer learning, achieving an accuracy of 95.44% and 96.19% on the ASTD and Aracust datasets, respectively.

Multiscale Change Detection Domain Adaptation Model Based on Illumination–Reflection Decoupling

In the change detection (CD) task, the substantial variation in feature distributions across different CD datasets significantly limits the reusability of supervised CD models. To alleviate this problem, we propose an illumination–reflection decoupled change detection multi-scale unsupervised domain adaptation model, referred to as IRD-CD-UDA. IRD-CD-UDA maintains its performance on the original dataset (source domain) and improves its performance on unlabeled datasets (target domain) through a novel CD-UDA structure and methodology. IRD-CD-UDA synergizes mid-level global feature marginal distribution domain alignment, classifier layer feature conditional distribution domain alignment, and an easy-to-hard sample selection strategy to increase the generalization performance of CD models on cross-domain datasets. Extensive experiments conducted on the LEVIR, SYSU, and GZ optical remote sensing image datasets demonstrate that the IRD-CD-UDA model effectively mitigates feature distribution discrepancies between source and target CD data, thereby achieving optimal recognition performance on unlabeled target domain datasets.

EasyDAM_V4: Guided-GAN-based cross-species data labeling for fruit detection with significant shape difference.

Traditional agriculture is gradually being combined with artificial intelligence technology. High-performance fruit detection technology is an important basic technology in the practical application of modern smart orchards and has great application value. At this stage, fruit detection models need to rely on a large number of labeled datasets to support the training and learning of detection models, resulting in higher manual labeling costs. Our previous work uses a generative adversarial network to translate the source domain to the target fruit images. Thus, automatic labeling is performed on the actual dataset in the target domain. However, the method still does not achieve satisfactory results for translating fruits with significant shape variance. Therefore, this study proposes an improved fruit automatic labeling method, EasyDAM_V4, which introduces the Across-CycleGAN fruit translation model to achieve spanning translation between phenotypic features such as fruit shape, texture, and color to reduce domain differences effectively. We validated the proposed method using pear fruit as the source domain and three fruits with large phenotypic differences, namely pitaya, eggplant, and cucumber, as the target domain. The results show that the EasyDAM_V4 method achieves substantial cross-fruit shape translation, and the average accuracy of labeling reached 87.8, 87.0, and 80.7% for the three types of target domain datasets, respectively. Therefore, this research method can improve the applicability of the automatic labeling process even if significant shape variance exists between the source and target domain.

Domain-interactive Contrastive Learning and Prototype-guided Self-training for Cross-domain Polyp Segmentation.

Accurate polyp segmentation plays a critical role from colonoscopy images in the diagnosis and treatment of colorectal cancer. While deep learning-based polyp segmentation models have made significant progress, they often suffer from performance degradation when applied to unseen target domain datasets collected from different imaging devices. To address this challenge, unsupervised domain adaptation (UDA) methods have gained attention by leveraging labeled source data and unlabeled target data to reduce the domain gap. However, existing UDA methods primarily focus on capturing class-wise representations, neglecting domain-wise representations. Additionally, uncertainty in pseudo labels could hinder the segmentation performance. To tackle these issues, we propose a novel Domain-interactive Contrastive Learning and Prototype-guided Self-training (DCL-PS) framework for cross-domain polyp segmentation. Specifically, domaininteractive contrastive learning (DCL) with a domain-mixed prototype updating strategy is proposed to discriminate class-wise feature representations across domains. Then, to enhance the feature extraction ability of the encoder, we present a contrastive learning-based cross-consistency training (CL-CCT) strategy, which is imposed on both the prototypes obtained by the outputs of the main decoder and perturbed auxiliary outputs. Furthermore, we propose a prototype-guided self-training (PS) strategy, which dynamically assigns a weight for each pixel during selftraining, filtering out unreliable pixels and improving the quality of pseudo-labels. Experimental results demonstrate the superiority of DCL-PS in improving polyp segmentation performance in the target domain. The code will be released at https://github.com/taozh2017/DCLPS.

Structured Domain Adaptation With Online Relation Regularization for Unsupervised Person Re-ID.

Unsupervised domain adaptation (UDA) aims at adapting the model trained on a labeled source-domain dataset to an unlabeled target-domain dataset. The task of UDA on open-set person reidentification (re-ID) is even more challenging as the identities (classes) do not have overlap between the two domains. One major research direction was based on domain translation, which, however, has fallen out of favor in recent years due to inferior performance compared with pseudo-label-based methods. We argue that domain translation has great potential on exploiting valuable source-domain data but the existing methods did not provide proper regularization on the translation process. Specifically, previous methods only focus on maintaining the identities of the translated images while ignoring the intersample relations during translation. To tackle the challenges, we propose an end-to-end structured domain adaptation framework with an online relation-consistency regularization term. During training, the person feature encoder is optimized to model intersample relations on-the-fly for supervising relation-consistency domain translation, which in turn improves the encoder with informative translated images. The encoder can be further improved with pseudo labels, where the source-to-target translated images with ground-truth identities and target-domain images with pseudo identities are jointly used for training. In the experiments, our proposed framework is shown to achieve state-of-the-art performance on multiple UDA tasks of person re-ID. With the synthetic→real translated images from our structured domain-translation network, we achieved second place in the Visual Domain Adaptation Challenge (VisDA) in 2020.

Self-ensembling with mask-boundary domain adaptation for optic disc and cup segmentation

Due to different retinal fundus image acquisition devices having various imaging principles, domain shift often occurs between different datasets. Hence, a segmentation network well-trained on one dataset (i.e., source domain) usually obtains very poor performance on another dataset (i.e., target domain), which results in us having to annotate the new dataset (target domain) to train the segmentation network again. However, annotating a new dataset is usually time-consuming and laborious. To address this problem, we proposed a novel unsupervised domain adaptation method for optic disc and cup segmentation. To be specific, we first utilized a domain adaptation method based on self-ensembling to effectively align the features of the source domain and target domain. Then, we designed a novel backbone network (MBU-Net) to make full use of the mask and boundary information to improve the segmentation performance of self-ensembling. Finally, we proposed an output-level adversarial domain adaptation (OADA) to address the domain shift problem of the structured output space in self-ensembling. In experiments, we test our proposed method on three different target domain datasets including Target Domain 1 (RIM-ONE_r3 dataset), Target Domain 2 (Drishti-GS dataset) and Target Domain 3 (REFUGE dataset). The experimental results demonstrate that our proposed method outperforms the compared state-of-the-art methods in the optic disc and cup segmentation tasks.

On Model-Based Transfer Learning Method for the Detection of Inter-Turn Short Circuit Faults in PMSM.

The early detection of an inter-turn short circuit (ITSC) fault is extremely critical for permanent magnet synchronous motors (PMSMs) because it can lead to catastrophic consequences. In this study, a model-based transfer learning method is developed for ITSC fault detection. The contribution can be summarized as two points. First of all, a Bayesian-optimized residual dilated CNN model was proposed for the pre-training of the method. The dilated convolution is utilized to extend the receptive domain of the model, the residual architecture is employed to surmount the degradation problems, and the Bayesian optimization method is launched to address the hyperparameters tuning issues. Secondly, a transfer learning framework and strategy are presented to settle the new target domain datasets after the pre-training of the proposed model. Furthermore, motor fault experiments are carried out to validate the effectiveness of the proposed method. Comparison with seven other methods indicates the performance and advantage of the proposed method.

Research on sentiment analysis method of opinion mining based on multi-model fusion transfer learning

With the popularity of social media, opinion mining has gradually become a popular research field. Among these fields, sentiment analysis mining is an important research direction in the field of opinion mining. It aims to reveal the public's sentiment tendency, and attitude towards specific topics or events by analyzing text data generated by users on online platforms and digital media. However, the large amount of opinion data usually lacks effective annotation, which limits the learning and construction of opinion models. Therefore, focusing on the problem of the scarcity of labeled data in opinion analysis, this paper proposes a mining method for public opinion sentiment analysis based on multi-model fusion transfer learning, that can make full use of the limited labeled data to improve the learning efficiency of sentiment features by integrating the advantages of different models. Additionally, it introduces a transfer learning strategy to enable the models of the target domains to perform better in the absence of labeled data. Furthermore, the attention mechanism is combined to strengthen the acquisition of key features of the emotional colors and improve the accuracy of sentiment analysis. Specifically, the paper uses the ERNIE model to generate dynamic representations of the text word vectors in the dataset. It also uses TextCNN and BiGRU to construct a joint model for extracting local and overall features of the text word vectors. The parameters of the feature layer of the trained model are migrated to the target domain through transfer learning. The attention mechanism is combined with the model to identify the extreme elements of the sentiment. Finally, the local and overall features are fused to achieve comprehensive mining of public opinion and emotional information. This method can effectively improve the accuracy and generalization of public opinion analysis in cases of data scarcity. In the experimental part, the paper conducts comparisons and analyses in eight aspects: word embedding model, model combination, attention mechanism, transfer learning, source domain dataset, target domain dataset, model training, and baseline model. The four indicators, namely accuracy, precision, recall, and F1-measure are used to evaluate the performance of the method. The experiments are thorough and detailed, demonstrating the effective improvement of opinion mining performance.