Feature Transfer Research Articles

An optimal deep learning approach for breast cancer detection and classification with pre-trained CNN-based feature learning mechanism

Breast cancer (BC) is the most dominant kind of cancer, which grows continuously and serves as the second highest cause of death for women worldwide. Early BC prediction helps decrease the BC mortality rate and improve treatment plans. Ultrasound is a popular and widely used imaging technique to detect BC at an earlier stage. Segmenting and classifying the tumors from ultrasound images is difficult. This paper proposes an optimal deep learning (DL)-based BC detection system with effective pre-trained transfer learning models-based segmentation and feature learning mechanisms. The proposed system comprises five phases: preprocessing, segmentation, feature learning, selection, and classification. Initially, the ultrasound images are collected from the breast ultrasound images (BUSI) dataset, and the preprocessing operations, such as noise removal using the Wiener filter and contrast enhancement using histogram equalization, are performed on the collected data to improve the dataset quality. Then, the segmentation of cancer-affected regions from the preprocessed data is done using a dilated convolution-based U-shaped network (DCUNet). The features are extracted or learned from the segmented images using spatial and channel attention including densely connected convolutional network-121 (SCADN-121). Afterwards, the system applies an enhanced cuckoo search optimization (ECSO) algorithm to select the features from the extracted feature set optimally. Finally, the ECSO-tuned long short-term memory (ECSO-LSTM) was utilized to classify BC into ‘3’ classes, such as normal, benign, and malignant. The experimental outcomes proved that the proposed system attains 99.86% accuracy for BC classification, which is superior to the existing state-of-the-art methods.

Exploring Latent Transferability of feature components

Feature disentanglement techniques have been widely employed to extract transferable (domain-invariant) features from non-transferable (domain-specific) features in Unsupervised Domain Adaptation (UDA). However, due to the complex interplay among high-dimensional features, the separated “non-transferable” features may still be partially informative. Suppressing or disregarding them, as commonly employed in previous methods, can overlook the inherent transferability. In this work, we introduce two concepts: Partially Transferable Class Features and Partially Transferable Domain Features (PTCF and PTDF), and propose a succinct feature disentanglement technique. Different with prior works, we don’t seek to thoroughly peel off the non-transferable features, as it is challenging practically. Instead, we take the two-stage strategy consisting of rough feature disentanglement and dynamic adjustment. We name our model as ELT because it can systematically Explore Latent Transferability of feature components. ELT can automatically evaluate the transferability of internal feature components, dynamically giving more attention to features with high transferability and less to features with low transferability, effectively solving the problem of negative transfer. Extensive experimental results have proved its efficiency. The code and supplementary file will be available at https://github.com/njtjmc/ELT.

MFFR-net: Multi-scale feature fusion and attentive recalibration network for deep neural speech enhancement

Deep neural networks (DNNs) have been successfully applied in advancing speech enhancement (SE), particularly in overcoming the challenges posed by nonstationary noisy backgrounds. In this context, multi-scale feature fusion and recalibration (MFFR) can improve speech enhancement performance by combining multi-scale and recalibrated features. This paper proposes a speech enhancement system that capitalizes on a large-scale pre-trained model, seamlessly fused with features attentively recalibrated using varying kernel sizes in convolutional layers. This process enables the SE system to capture features across diverse scales, enhancing its overall performance. The proposed SE system uses a transferable features extractor architecture and integrates with multi-scaled attentively recalibrated features. Utilizing 2D-convolutional layers, the convolutional encoder-decoder extracts both local and contextual features from speech signals. To capture long-term temporal dependencies, a bidirectional simple recurrent unit (BSRU) serves as a bottleneck layer positioned between the encoder and decoder. The experiments are conducted on three publicly available datasets including Texas Instruments/Massachusetts Institute of Technology (TIMIT), LibriSpeech, and Voice Cloning Toolkit+Diverse Environments Multi-channel Acoustic Noise Database (VCTK+DEMAND). The experimental results show that the proposed SE system performs better than several recent approaches on the Short-Time Objective Intelligibility (STOI) and Perceptual Evaluation of Speech Quality (PESQ) evaluation metrics. On the TIMIT dataset, the proposed system showcases a considerable improvement in STOI (17.3%) and PESQ (0.74) over the noisy mixture. The evaluation on the LibriSpeech dataset yields results with a 17.6% and 0.87 improvement in STOI and PESQ.

Theories and Methods for Indoor Positioning Systems: A Comparative Analysis, Challenges, and Prospective Measures.

In the era of the Internet of Things (IoT), the demand for accurate positioning services has become increasingly critical, as location-based services (LBSs) depend on users' location data to deliver contextual functionalities. While the Global Positioning System (GPS) is widely regarded as the standard for outdoor localization due to its reliability and comprehensive coverage, its effectiveness in indoor positioning systems (IPSs) is limited by the inherent complexity of indoor environments. This paper examines the various measurement techniques and technological solutions that address the unique challenges posed by indoor environments. We specifically focus on three key aspects: (i) a comparative analysis of the different wireless technologies proposed for IPSs based on various methodologies, (ii) the challenges of IPSs, and (iii) forward-looking strategies for future research. In particular, we provide an in-depth evaluation of current IPSs, assessing them through multidimensional matrices that capture diverse architectural and design considerations, as well as evaluation metrics established in the literature. We further examine the challenges that impede the widespread deployment of IPSs and highlight the potential risk that these systems may not be recognized with a single, universally accepted standard method, unlike GPS for outdoor localization, which serves as the golden standard for positioning. Moreover, we outline several promising approaches that could address the existing challenges of IPSs. These include the application of transfer learning, feature engineering, data fusion, multisensory technologies, hybrid techniques, and ensemble learning methods, all of which hold the potential to significantly enhance the accuracy and reliability of IPSs. By leveraging these advanced methodologies, we aim to improve the overall performance of IPSs, thus paving the way for more robust and dependable LBSs in indoor environments.

ADA-UDA: A transferable transformer framework for rumor detection using Adversarial Domain Alignment within Unsupervised Domain Adaptation

The rapid spread of misinformation on social media has posed significant challenges, particularly in the early detection of rumors, which is critical to mitigating their negative impact. However, the limited data available during the initial stages of trending topics poses significant challenges, including analysis constraints and model overfitting. Therefore, the adoption of specific methods are needed to identify patterns from historical sample data and enable effective knowledge transfer, facilitating inductive reasoning with limited data on new topics. To address these challenges, we propose an Adversarial Domain Alignment Unsupervised Domain Adaptation (ADA-UDA) method based on the Transformer architecture. Our approach leverages labeled historical data from the source domain alongside a limited quantity of unlabeled data from new trending topics in the target domain. At the heart of our method is the Parameter Transferable Module (PTM), which guides the Transformer to focus on transferable and distinguishable features, thereby enhancing the model’s ability to perform effective inductive reasoning with limited data. We conducted extensive experiments to evaluate our method, benchmarking it against current mainstream rumor detection techniques. The findings indicate that our ADA-UDA method outperforms existing approaches, underscoring its potential for early and accurate rumor detection in emerging topics.

A Novel Fault Diagnosis Method for Rolling Bearings Based on Multi-Domain Adaptation Neural Network

In recent years, data-driven approaches have shown promise in fault diagnosis. However, in practical industrial applications, challenges such as discrepancies between source domain and target domain data distributions, coupled with limited labeled fault data, have hindered the performance of traditional domain adaptation algorithms in bearing fault diagnosis. To address this issue, this study introduces a novel method based on MDANN (Multi Domain Adaptation Neural Network) for diagnosing rolling bearing faults using unlabeled data. Initially, the raw vibration signals are preprocessed using wavelet packet decomposition and reconstruction (WPT), which helps minimize signal redundancy while preserving critical features. Following this, the multi-kernel maximum mean discrepancy (MK-MMD) algorithm is employed to compute the differences in input feature values, and the MDANN network parameters are adjusted via backpropagation, allowing the network to extract domain-invariant features. To ensure the unlabeled target domain data can be effectively used in training, a pseudo-labeling strategy is applied, where the label with the highest probability is treated as the true label, thereby enabling the model to learn from the target domain data and improve the acquisition of reliable diagnostic knowledge. The method is validated using two publicly available datasets, CWRU and PU, with experimental results demonstrating that it outperforms conventional domain adaptation techniques in terms of diagnostic accuracy. This indicates the proposed approach is highly effective in capturing transferable features and addressing the distributional differences between datasets.

A bearing fault diagnosis method for unknown operating conditions based on differentiated feature extraction

Under unknown operating conditions, the domain generalization approach based on domain metrics is commonly used for rolling bearing fault diagnostics. Nevertheless, in the event of equipment failure under unknown operating conditions, focusing solely on the transferable characteristics across domains may result in the unintentional neglect of domain-specific features. To address the problems mentioned, the present study introduces a feature decomposition learning method that simultaneously extracts inter-domain transferable and domain-specific features. This method aims to obtain richer feature information by constructing different feature extractors. For the extraction of transferable features, a joint metric method based on central moment differences is devised. A difference maximization method is employed to extract domain-specific features. The experimental findings demonstrate that the proposed technique exhibits greater defect detection capacity across two datasets.

A New N‐functionalized Pyrrole Derivative Bearing Ruthenium Tris(2,2’‐bipyridine) Subunit

AbstractIn recent years, nitrogen‐based ruthenium complexes have become an important material for dye‐sensitized solar cells and photo‐catalytic H2O and CO2 reduction research and applications, thanks to their charge transfer feature from the metal to the ligand. The electronic and spectral properties of such materials can be easily modified by changing the substituents and ligands. In this study, a new ruthenium (II) complex with N‐functionalized pyrrole derivative ([Ru(bpy)2(bpyCONHArPy)]+2 complex) was successfully synthesized and characterized. Its optical and electrochemical properties were studied by means of fluorescence, UV‐vis spectrometry, FTIR spectroscopy and cyclic voltammetry. These results were compared with that of its components (4‐(1H‐pyrrol‐1‐yl) aniline, N4,N4′‐bis(4‐(1H‐pyrrol‐1‐yl)phenyl)‐[2,2′‐bipyridine]‐4,4′‐dicarboxamide and ligand [Ru(bpy)3]+2). Related theoretical studies were also studied to find energy levels of complex. The present work is carried out to understand what chemical changes bring when combination of all the pieces of the puzzle into a chemical structure step by step.

A Feature of the Horizontal Directional Solidification (HDS) Method Affects the Microstructure of Al2O3/YAG Eutectic Ceramics

The solidification processes of two compositions, hypereutectic (21.0 mol% Y2O3–79.0 mol% Al2O3) and eutectic (18.5 mol% Y2O3–81.5 mol% Al2O3), were used via the horizontal directional solidification (HDS) method to produce two ingots with dimensions of 317 × 220 × 35 mm and 210 × 180 × 35 mm, respectively. The first ingot was heterogeneous and characterized by a two-layer structure with an expressed horizontal boundary, which is parallel to the solidification direction (an experimental fact observed for the first time), separating eutectic-type ceramics in the upper layer from the lower one containing the YAG dendrites. Considering the heat transfer feature characteristic of the HDS method and its action during the solidification of materials scattering thermal radiation, an explanation of the occurrence of such structure has been proposed. On this basis, the solidification parameters of the second ingot, providing its homogeneous structure, were selected. Characterization of the crystallographic texture and microstructure of both ingots revealed the advantage of the second solidification processing conditions.

Wave masked autoencoder: An electrocardiogram signal diagnosis model based on wave making strategy

ECG signal labeling is time-consuming and laborious, significantly constraining the performance of deep learning models on ECG diagnosis. Self-supervised learning methods offer a means to mitigate the reliance on annotated ECG data. Nevertheless, the direct transposition of these methods to ECG diagnostics may encounter challenges stemming from deficient domain knowledge and incongruous model architectures. Therefore, a self-supervised learning model that introduces ECG prior knowledge, called wave masked autoencoder (WMAE), is proposed. Firstly, considering that various waves that make up the ECG signal are key features for diagnosing arrhythmias, we design a wave masking strategy as an auxiliary task for self-supervised learning to introduce ECG-related knowledge. Secondly, sliding window slicing is utilized to divide each ECG signal into a series of subsequences, which are mapped into deep features with a convolutional feature mapper to increase the information density of basic semantic units and learn complete waveform representations. In this way, the proposed WMAE can capture robust and transferable ECG signal features from unlabeled data, thereby alleviating the problem of insufficient annotated ECG data and improving the performance of the model on cardiovascular disease diagnosis. Experimental results show that WMAE is significantly better than other benchmark models and has high practical value.

A crop disease severity index derived from transfer learning and feature fusion using enhanced OPTICS algorithm

The adoption of automated methods for the identification and assessment of tomato-related disorders is highly sought-after in the agriculture sector. Using this technology is crucial for reducing wasteful spending, increasing the efficiency of treatments, and ultimately growing more resilient crops by reducing losses in agricultural output and maximising the effectiveness of these processes. An automated method has been suggested for accurately identifying and classifying diseases using a single photograph. The described method for disease detection in tomato plants makes use of a computer vision-based technique. Image processing, ML, and deep learning are just a few of the methods that this strategy uses. The goal of this approach is to prevent tomato crops from being damaged by various illnesses by reducing the need of conventional procedures. Bacterial spot, early blight, late blight, leaf mould, spider mites, target spot, spotted spider mite, mosaic virus, and yellow leaf curl are all examples of these illnesses. The following ten diseases frequently strike tomato crops in India. By utilising picture segmentation in combination with the Enhanced OPTICS algorithm (EOPTICSA), the affected area of the tomato plant may be precisely detected and defined after image pre-processing procedures have been used. It may be necessary to look for certain visual signs in order to diagnose the previously mentioned illnesses. The primary goal of this study was to evaluate the efficacy of the EOPTICSA method for detecting diseases in plant leaves. To eliminate the geometric features associated with colour, texture, and leaf arrangement in the provided plant pictures, image segmentation and edge detection methods are employed. Using these methods allows us to achieve our goal. Various efficacy measures are used to assess and provide a technique recommendation. This research shows that when performance metrics are used to implement these strategies, the suggested strategy outperforms the current methods in terms of accuracy, precision, and F1-score. The process of detecting sickness involves several consecutive steps. Capturing images, segmenting them, detecting edges, and determining the infection’s severity are all steps in this process. To accomplish the goal of recognising and categorising different types of diseases that might impact tomato plants, the method of transfer learning is employed. As soon as the problem is identified, it is recommended to take proactive measures to help individuals and organisations involved in agriculture address the effects of these disorders using appropriate measures.

From processing to properties: Enhancing machine learning models with microstructural information in polymer nanocomposites

For polymers and their composites, processing conditions and the resultant microstructures are crucial in determining their properties. Traditional machine learning (ML) methods typically focus on establishing direct relationships between processing parameters and material properties, often overlooking the critical intermediate step of how processing influences microstructure, limiting the predictive accuracy. In this study, we introduce an approach that first establishes a detailed relationship between processing parameters and the resultant microstructure, and then uses transfer learning and feature fusion to integrate this relationship into the prediction of material properties. Using carbon black-reinforced rubber composites (CRC) as an example, we compared ML models in predicting mechanical properties from processing data. A multi-task deep neural network performed best achieving an R2 of 0.763 with only processing data as input. When incorporating transfer learning and feature fusion, the R2 improved to 0.852 and 0.878, respectively. Shapley explanation analysis validated our approach, highlighting the importance of integrating processing, microstructure, and properties in ML models. This method emphasizes the critical effect of comprehensively considering processing–(micro)structure–property (P–S–P) relationships, leading to more accurate predictions in polymer composite research.

Kernel Extreme Learning Machine with Discriminative Transfer Feature and Instance Selection for Unsupervised Domain Adaptation

The goal of domain adaptation (DA) is to develop a robust decision model on the source domain effectively generalize to the target domain data. State-of-the-art domain adaptation methods typically focus on finding an optimal inter-domain invariant feature representation or helpful instances from the source domain. In this paper, we propose a kernel extreme learning machine with discriminative transfer features and instance selection (KELM-DTF-IS) for unsupervised domain adaptation tasks, which consists of two steps: discriminative transfer feature extraction and classification with instance selection. At the feature extraction stage, we extend cross domain mean approximation(CDMA) by incorporating a penalty term and develop discriminative cross domain mean approximation (d-CDMA) to optimize the category separability between instances. Subsequently, d-CDMA is integrated into kernel ELM-AutoEncoder(KELM-AE) for extracting inter-domain invariant features. During the classification process, our approach uses CDMA metrics to compute a weights to each source instances based on their impact in reducing distribution differences between domains. Instances with a greater effect receive higher weights and vice versa. These weights are then used to distinguish and select source domain instances before incorporating them into weight KELM for proposing an adaptive classifier. Finally, we apply our approach to conduct classification experiments on publicly available domain adaptation datasets, and the results demonstrate its superiority over KELM and numerous other domain adaptation approaches.

Photo-isomerization enabled reversible wavelength switching in fiber random laser for color image encryption.

Random lasers owing the functionality of generating random spectra facilitate the chaotic encrypted systems essential for cryptography in the current information epoch. Nevertheless, single wavelength bands of random lasers provide an unsuitable key for image encryption that causes outline interpretation and a fragile complex dual chaotic encryption demanding secured image encryption. This research presents an inevitable development of a reversible switchable wavelength fiber random laser composed of the mixture of highly polarized intramolecular charge transfer dye molecules and the optimum concentration of titanium dioxide acting as gain and efficient scattering mediums respectively within a polyvinyl alcohol matrix. This mixture with a certain ratio is coated on a fiber employing a dip coated method, followed by a layer of polydimethylsiloxane to facilitate with high coefficient of thermal expansion. Random laser emission is enabled with dynamically switchable wavelengths obeying the excited state intramolecular proton transfer phenomenon under the photo-isomerization. The optimum scatters concentration yields a lower threshold of 32 µJ/cm2 with full width at half maximum of 0.4 nm and dual emission reversible switchable wavelength bands centered around 443 nm and 464 nm attributed to inter charge transfer feature of the dye molecules. Thereby, the dual reversible switchable wavelength bands feed as input for a dual chaotic color image encryption system. Further, in this integrated system, beam divergence of random laser emissions remains less than 20° during both situations of with- and without irradiation. This delicate approach paves the way in laying the foundation about the applicability of fiber random lasers in an information security system.

Bilaterally Normalized Scale-Consistent Sinkhorn Distance for Few-Shot Image Classification.

Few-shot image classification aims at exploring transferable features from base classes to recognize images of the unseen novel classes with only a few labeled images. Existing methods usually compare the support features and query features, which are implemented by either matching the global feature vectors or matching the local feature maps at the same position. However, few labeled images fail to capture all the diverse context and intraclass variations, leading to mismatch issues for existing methods. On one hand, due to the misaligned position and cluttered background, existing methods suffer from the object mismatch issue. On the other hand, due to the scale inconsistency between images, existing methods suffer from the scale mismatch issue. In this article, we propose the bilaterally normalized scale-consistent Sinkhorn distance (BSSD) to solve these issues. First, instead of same-position matching, we use the Sinkhorn distance to find an optimal matching between images, mitigating the object mismatch caused by misaligned position. Meanwhile, we propose the intraimage and interimage attentions as the bilateral normalization on the Sinkhorn distance to suppress the object mismatch caused by background clutter. Second, local feature maps are enhanced with the multiscale pooling strategy, making the Sinkhorn distance possible to find a consistent matching scale between images. Experimental results show the effectiveness of the proposed approach, and we achieve the state-of-the-art on three few-shot benchmarks.

A robust domain distribution alignment discriminative network driven by physical samples for rotor-bearing fault diagnosis

The diagnostic performance of most intelligent models depends on having a sufficient number of labeled samples, which are difficult to obtain in real-world scenarios. Physics-guided sample generation emerges as a promising remedy for addressing the scarcity of labeled samples during model training. However, existing studies have tended to overlook the coupling effect between the marginal and conditional distributions of physical and real samples, leading to domain confusion and a subsequent reduction in model generalization and diagnostic accuracy. Therefore, this study introduces a novel approach: the physically driven sample-based domain distribution alignment discriminative network (DDADN), which integrates dynamic modeling and domain adaptation. Specifically, a lumped parameter model is established to generate complete defect samples, subsequently optimized using the Pearson correlation coefficient (PCC). Next, an enhanced joint distribution adaptation (EJDA) mechanism is employed to harmonize distributions across domains, effectively aligning both overall and class-specific distributions within the domains and thus achieving adaptive domain confusion. Meanwhile, to ensure robust feature classification and extraction, an adaptive softmax (A-softmax) loss function is introduced. Finally, in unlabeled real operating conditions, the proposed multitask scheme adeptly transfers knowledge between physical and real signals, yielding impressive classification accuracies of 97.78% and 98.34% for the experimental and Case Western Reserve University (CWRU) datasets, respectively. These results underscore the superiority of the proposed scheme in terms of learning transferable features and maintaining high diagnostic accuracy and generalization capability.

Hybrid Deep Transfer Learning and Feature Fusion Architecture for Diabetic Retinopathy Classification and Severity Grading

Diabetic Retinopathy (DR) is the leading cause of blindness among individuals with diabetes. Automating the diagnosis of DR has the potential to greatly benefit patients by enabling early detection and intervention, thus reducing the risk of blindness. The primary objective of this research is to develop a robust approach for the classification of DR and to analyze its severity grading. By achieving this, we aim to provide an effective tool for accurate diagnosis and assessment of DR, contributing to improved patient care and outcomes. The current literature review analysis reported the importance of deep learning in computer vision based applications. Moreover, plenty of pre-trained models are also present which can be used for classification tasks. Therefore, we present a hybrid DL classification approach by combining Inception V3, VGG-19 and ResNet 50. The proposed architecture uses transfer learning, and feature fusion model to produce the weighted feature vector which is used for classification analysis. The proposed approach is experimented on publicly available datasets APTOS-2019 and Messidor. The performance is measured in terms of accuracy, precision, recall and F1-score.

Deep conditional adversarial subdomain adaptation network for unsupervised mechanical fault diagnosis

In recent years, intelligent fault diagnosis technology has rapidly advanced, achieving remarkable results. Most of them assume that the source and target domains have similar distributions. However, the actual working conditions of rotating machinery are variable, leading to a significant distribution discrepancy between the source and target domains. The majority of existing domain adaptation methods mainly consider either the marginal distribution or the conditional distribution between the source and target domains, which limits the improvement of diagnostic performance in the target domain. Considering the discrepancy in domain distributions, a deep conditional adversarial subdomain adaptation network (DCASAN) is proposed in this paper for unsupervised mechanical fault diagnosis under unknown working conditions. In DCASAN, a backbone network based on the feature fusion of convolutional neural network (CNN) and graph convolutional network (GCN) is designed for feature extraction. Furthermore, a conditional adversarial subdomain adaptation strategy is proposed, which simultaneously considers both the marginal distribution and the conditional distribution. Under the joint constraints of marginal and conditional distribution alignment, the extracted features exhibit improved inter-class discriminability and intra-class consistency, effectively reducing the distribution discrepancy between different domains. Results from multiple transfer tasks on three rotating machinery datasets demonstrate that the proposed method can learn rich transferable features. The overall average accuracy of the proposed method on these three datasets is 88.07%, 89.17%, and 96.63%, respectively, significantly outperforming other methods in terms of robustness and generalization.

Pemanfaatan Fitur Shopee Sebagai Media Silaturahmi

This research is entitled "Utilization of Shopee Features as Friendshio Media". The rapid development of technology affects changes in the new communication style carried out by the community. With the development of this technology, online media is not only a means of communicating but also used as a fulfillment of individual needs. With this, current technological advances can change patterns in staying in touch between individuals efficiently. The purpose of this study is to determine the utilization of Shopee features as a medium for friendship. This research uses qualitative research methods and uses data collection techniques by means of observation and interviews. This study uses the Uses and Gratifications theory, the results of which show that the utilization of the THR Shopee feature and the Bank Transfer feature is an innovation that supports many activities between individuals to stay in touch in different ways, informants use this feature with their various needs. There are 5 dimensions of needs that are in accordance with the informants in the Uses and Gratifications theory in utilizing the Shopee feature as a gathering media, namely: cognitive needs related to information and knowledge of the surrounding environment, affective needs related to the pleasant experience of individuals in utilizing a medium, personal needs related to work or personality, social needs related to family or friends in the neighborhood, and tension release needs related to the individual's desire to release pressure after doing something.

SDP-MTF: A Composite Transfer Learning and Feature Fusion for Cross-Project Software Defect Prediction

Software defect prediction is critical for improving software quality and reducing maintenance costs. In recent years, Cross-Project software defect prediction has garnered significant attention from researchers. This approach leverages transfer learning to apply the knowledge from existing projects to new ones, thereby enhancing the universality of predictive models. It provides an effective solution for projects with limited historical defect data. Nevertheless, current methodologies face two main challenges: first, the inadequacy of feature information mining, where code statistical information or semantic information is used in isolation, ignoring the benefits of their integration; second, the substantial feature disparity between different projects, which can lead to insufficient effect during transfer learning, necessitating additional efforts to narrow this gap to improve precision. Addressing these challenges, this paper proposes a novel methodology, SDP-MTF (Software Defect Prediction using Multi-stage Transfer learning and Feature fusion), that combines code statistical features, deep semantic features, and multiple feature transfer learning methods to enhance the predictive effect. The SDP-MTF method was empirically tested on single-source cross-project software defect prediction across six projects from the PROMISE dataset, benchmarked against five baseline algorithms that employ distinct features and transfer methodologies. Our findings indicate that SDP-MTF significantly outperforms five classical baseline algorithms, improving the F1-Score by 8% to 15.2%, thereby substantively advancing the precision of cross-project software defect prediction.