Applying Transfer Learning Research Articles

Apple Flower Recognition Using Convolutional Neural Networks with Transfer Learning and Data Augmentation Technique

Automated monitoring of apple flowers using convolutional neural networks will enable informed decision-making for planning thinning and fruit set operations, optimizing crop load, preventing fruiting periodicity, and enhancing crop quality. The article presents the results of apple flower recognition quality on images using the YOLOv8 (You Only Look Once version 8) convolutional neural network model with the application of transfer learning and data augmentation technique. Pre-trained weights on the Common Objects in Context (COCO) dataset were utilized in the research. To expand the dataset and enhance model performance, the tools Flip, 90° Rotate, Crop, Rotation, Shear, Grayscale, Hue, Saturation, Brightness, Exposure, Blur, Noise, and Cutout were applied. The result showed that artificial augmentation of the training dataset significantly improves the quality of training for the YOLOv8 convolutional neural network model, increasing the average accuracy of detecting class features apple flowers. The analysis of the Precision-Recall curve allowed establishing a classification threshold (0.47) that provides the optimal balance between precision and recall in recognizing apple flowers at the flowering stage in images. The mAP metric for recognizing the «flower» class (flowers in the flowering stage) was 0.595. The analysis of the obtained results revealed an increase in the Precision metric by 2.1%, Recall metric by 10.13%, and mAP@0.5 metric by 5.31% when using the augmentation technique. The obtained results indicate a significant improvement in the performance of the model in recognizing apple flowers when applying the augmentation technique to the training dataset.

MedTransCluster: Transfer learning for deep medical image clustering

This work introduces the “MedTransCluster” framework, a novel approach to medical image clustering in chest radiography through the application of transfer learning, leveraging the capabilities of pre-trained deep learning models. Our evaluation encompassed a variety of neural networks, considering their adaptability to the nuances of medical imaging data. The study incorporated four renowned clustering algorithms and an expanded set of evaluation metrics, offering a comprehensive comparison and a refined analysis of these models’ ability to cluster complex diagnostic features. Notably, EfficientNetB0 coupled with DBSCAN clustering algorithm achieved a silhouette score of 0.924131, and ResNet152 with KMeans displayed a Calinski Harabasz score of 9655.213964, indicating their superior proficiency in capturing the intricacies of medical features. These results emphasize the critical importance of model refinement within the healthcare imaging sphere and underscore the potential of methodologies like MedTransCluster in enhancing diagnostic accuracy and patient outcomes.

Optimizing postprandial glucose prediction through integration of diet and exercise: Leveraging transfer learning with imbalanced patient data.

In recent years, numerous methods have been introduced to predict glucose levels using machine-learning techniques on patients' daily behavioral and continuous glucose data. Nevertheless, a definitive consensus remains elusive regarding modeling the combined effects of diet and exercise for optimal glucose prediction. A notable challenge is the propensity for observational patient datasets from uncontrolled environments to overfit due to skewed feature distributions of target behaviors; for instance, diabetic patients seldom engage in high-intensity exercise post-meal. In this study, we introduce a unique application of Bayesian transfer learning for postprandial glucose prediction using randomized controlled trial (RCT) data. The data comprises a time series of three key variables: continuous glucose levels, exercise expenditure, and carbohydrate intake. For building the optimal model to predict postprandial glucose levels we initially gathered balanced training data from RCTs on healthy participants by randomizing behavioral conditions. Subsequently, we pretrained the model's parameter distribution using RCT data from the healthy cohort. This pretrained distribution was then adjusted, transferred, and utilized to determine the model parameters for each patient. The efficacy of the proposed method was appraised using data from 68 gestational diabetes mellitus (GDM) patients in uncontrolled settings. The evaluation underscored the enhanced performance attained through our method. Furthermore, when modeling the joint impact of diet and exercise, the synergetic model proved more precise than its additive counterpart. An innovative application of the transfer-learning utilizing randomized controlled trial data can improve the challenging modeling task of postprandial glucose prediction for GDM patients, integrating both dietary and exercise behaviors. For more accurate prediction, future research should focus on incorporating the long-term effects of exercise and other glycemic-related factors such as stress, sleep.

Application of Transfer Learning in Smart Agriculture to Combat Black Rot Bacteria

With diseases and weeds alone accounting for approximately 30% of productivity losses in agriculture worldwide, plant diseases are a significant factor. Farmers frequently consult experts to diagnose their crops in an effort to reduce these losses. On the other hand, conventional visual assessment techniques focused on elements like leaf color, shape, and texture can be inefficient, time-consuming, and expensive. This study suggests using symptom traits like yellow leaves, black lesions, wilting, and plant mortality in crops like apple and grape to anticipate plant diseases, notably black rot. AlexNet, VGGNet, ResNet, and Inception are the four deep learning models examined in the study; VGGNet achieved the greatest accuracy of 98.19%.

Harnessing Deep Learning Techniques for Enhanced Detection and Classification of Cracks in Pavement Imagery

With the increased demand for infrastructure maintenance and the evolving capabilities of deep learning, there is a pressing need to apply advanced techniques for efficient and accurate detection of structural anomalies. This paper delves into the application of Convolutional Neural Networks (CNNs) and transfer learning to address the problem of pavement crack classification using images. Three distinctive models were explored: a 4-layer CNN, a 2-layer CNN, and a VGG-based transfer learning model. Through comprehensive experimentation, we demonstrate the efficacy of these models in accurately classifying Pavement cracks from a curated dataset. Our findings indicate that while each model possesses its unique strengths, the VGG-based transfer learning model exhibits superior performance in terms of precision and recall. This research not only contributes to the growing body of knowledge in infrastructure maintenance using deep learning but also provides practical insights for professionals aiming to employ automated systems for Pavement inspection.

An ensemble of ResNet model for classification of Parkinson disease

Parkinson disease (PD) is one of the neurological illnesses incurred. However, there is a no chance to recognize PD. A fine motor symptom has been identified in this study. A group of patients with PD, as well as the healthy group, is used in the research. The authors have developed a technique that can determine whether a patient has PD or not. Using deep learning methods, the same design generalizing neural networks in the brain can be solved. The categorization of patients with PD and non-PD behavior is found from the analysis of spiral and wave forms using CNN model. Various CNN models were used in the experiment by transfer learning and spiral and wave data sketches. With the help of spiral sketching, the system achieved an accuracy of 96.67% using the ResNet50 model. The main objective of this paper is to explore the application of transfer learning, which improved the performance of the model.

Mordo2: A Personalization Framework for Silent Command Recognition.

Wearable human-computer interactions in daily life are increasingly encouraged by the prevalence of intelligent wearables. It poses a demanding requirement of micro-interaction and minimizing social awkwardness. Our previous work demonstrated the feasibility of recognizing silent commands through around-ear biosensors with the limitation of user adaptation. In this work, we ease the limitation by a personalization framework that integrates spectral factorization of signals, temporal confidence rejection and commonly used transfer learning algorithms. Specifically, we first empirically formulate the user adaptation issue by presenting the accuracies of applying transfer learning algorithms to our previous method. Second, we improve the signal-to-noise ratio by proposing the supervised spectral factorization method that learns the amplitude and phase mappings between around-ear signals and the signals of articulated facial muscles. Third, we leverage the time continuity of commands and introduce the time decay into confidence rejection. Finally, extensive experiments have been conducted to evaluate the feasibility and improvements. The results indicate an average accuracy of 92.38% which is significantly larger than solely using transfer learning algorithms. And a comparable accuracy can be achieved with significantly reduced data of new users. The overall performance shows the framework can significantly improve the accuracy of user adaptations. The work would aid a further step toward commercial products for silent command recognition and inspire the solution to the user adaptation challenge of wearable human-computer interactions.

Models and means of clothing elements patterns classification using machine learning

The task of pattern classification remains relevant in the fields of trends, style, fashion, personalization, manufacturing, and design. Research aimed at the design and development of models and means of classification of patterns of clothing elements using machine learning is highlighted. The study addresses a pertinent issue in computer vision, namely: increasing the efficiency of classification of patterns of clothing elements. The research was conducted with a proprietary dataset comprising 600 images. The following patterns are defined for classification: “checkered”, “dotted”, “vegetation/floral”, “print”, “solid”, “striped”. A convolutional neural network was developed using the Python programming language and deep learning frameworks Keras and TensorFlow. The scalable Keras-Tuner framework was used to optimize the hyperparameters of the developed network. The structure of the convolutional neural network includes an input layer, a feature extraction part, and a pattern type determination part. The architecture of the applied convolutional neural network is described. The CUDA Toolkit, the cuDNN library and the WSL layer are applied to train a convolutional neural network using a GPU, significantly speeding up the training process. Metrics including accuracy, precision, and recall were used to evaluate the developed convolutional neural network. The web application is developed in the Python programming language with the FastAPI framework. The web application has a described API for interacting with a convolutional neural network, and uses the Pillow (PIL) libraries for working with images and Rembg for image background removal. The user interface is developed in the JavaScript programming language with HTML, CSS and the React framework. The user interface is presented as an intuitive tool for interacting with the system. The developed software uses the modular principle, which allows for rapid modernization of the software. As a result of applying transfer learning, a testing accuracy of 93.33% was achieved, and with fine-tuning, the final version of the convolutional neural network for the classification of patterns of clothing elements with a test accuracy of 95% was obtained. The trained neural network was tested on new images of the specified types of patterns, examples for two patterns are given.

Simultaneous EEG-fNIRS Data Classification Through Selective Channel Representation and Spectrogram Imaging.

The integration of electroencephalography (EEG) and functional near-infrared spectroscopy (fNIRS) can facilitate the advancement of brain-computer interfaces (BCIs). However, existing research in this domain has grappled with the challenge of the efficient selection of features, resulting in the underutilization of the temporal richness of EEG and the spatial specificity of fNIRS data.To effectively address this challenge, this study proposed a deep learning architecture called the multimodal DenseNet fusion (MDNF) model that was trained on two-dimensional (2D) EEG data images, leveraging advanced feature extraction techniques. The model transformed EEG data into 2D images using a short-time Fourier transform, applied transfer learning to extract discriminative features, and consequently integrated them with fNIRS-derived spectral entropy features. This approach aimed to bridge existing gaps in EEG-fNIRS-based BCI research by enhancing classification accuracy and versatility across various cognitive and motor imagery tasks.Experimental results on two public datasets demonstrated the superiority of our model over existing state-of-the-art methods.Thus, the high accuracy and precise feature utilization of the MDNF model demonstrates the potential in clinical applications for neurodiagnostics and rehabilitation, thereby paving the method for patient-specific therapeutic strategies.

Revolutionizing Breast Cancer Detection: Enhanced Mammogram Analysis with Modified Xception and Self-Attention

In the realm of precision medicine, the intersection of state-of-the-art technology and disease identification has ushered in a new era of accuracy in healthcare diagnostics. This paper reviews a groundbreaking research endeavor aiming to enhance breast cancer detection in mammograms through the synergistic application of transfer learning, architectural modifications, and advanced optimization techniques.
 At the core of this research is the concept of transfer learning, leveraging insights gained from one domain to illuminate another. The study extends this paradigm to the Xception architecture, renowned for its proficiency in discerning intricate patterns within images. However, the novelty lies in the intelligent modification of Xception to be specifically attuned to breast cancer detection. By fine-tuning the network's final layers, the model's innate ability to understand features is harmonized with the complexities of mammographic images, ensuring sensitivity to nuanced markers of potential malignancy.
 A distinctive aspect of this research is the incorporation of a self-attention mechanism, mirroring human visual processing. This mechanism dynamically highlights crucial regions within mammograms, transforming the model into an active interpreter capable of identifying subtle textures, shapes, and edges indicative of breast cancer. This adaptive approach enhances the model's finesse in navigating diverse breast cancer cases effectively.
 Throughout the training phase, optimization is pivotal for steering the research towards success. The utilization of the Adam optimizer, known for its adaptability in learning rates and moment estimations, guides the process, ensuring precise gradient descent through complex patterns. The integration of the rectified linear unit (ReLU) activation function further empowers the model to capture intricate relationships within data, enhancing its ability to identify subtle cancer markers. This review comprehensively explores the innovative strides made in breast cancer detection, shedding light on the nuanced interplay of transfer learning, architectural adjustments, self-attention mechanisms, and advanced optimization techniques. The modified Xception model emerges as a promising tool in the pursuit of accurate and sensitive breast cancer diagnostics.

Exploring the Use of Contrastive Language-Image Pre-Training for Human Posture Classification: Insights from Yoga Pose Analysis

Accurate human posture classification in images and videos is crucial for automated applications across various fields, including work safety, physical rehabilitation, sports training, or daily assisted living. Recently, multimodal learning methods, such as Contrastive Language-Image Pretraining (CLIP), have advanced significantly in jointly understanding images and text. This study aims to assess the effectiveness of CLIP in classifying human postures, focusing on its application in yoga. Despite the initial limitations of the zero-shot approach, applying transfer learning on 15,301 images (real and synthetic) with 82 classes has shown promising results. The article describes the full procedure for fine-tuning, including the choice for image description syntax, models and hyperparameters adjustment. The fine-tuned CLIP model, tested on 3826 images, achieves an accuracy of over 85%, surpassing the current state-of-the-art of previous works on the same dataset by approximately 6%, its training time being 3.5 times lower than what is needed to fine-tune a YOLOv8-based model. For more application-oriented scenarios, with smaller datasets of six postures each, containing 1301 and 401 training images, the fine-tuned models attain an accuracy of 98.8% and 99.1%, respectively. Furthermore, our experiments indicate that training with as few as 20 images per pose can yield around 90% accuracy in a six-class dataset. This study demonstrates that this multimodal technique can be effectively used for yoga pose classification, and possibly for human posture classification, in general. Additionally, CLIP inference time (around 7 ms) supports that the model can be integrated into automated systems for posture evaluation, e.g., for developing a real-time personal yoga assistant for performance assessment.

Analysis of the Machine-Specific Behavior of Injection Molding Machines.

The performance of an injection molding machine (IMM) influences the process and the quality of the parts manufactured. Despite increasing data collection capabilities, their machine-specific behavior has not been extensively studied. To close corresponding research gaps, the machine-specific behavior of two hydraulic IMMs of different sizes and one electric IMM were compared with each other as part of the investigations. Both the start-up behavior from the cold state and the behavior of the machine at different operating points were considered. To complement this, the influence of various material properties on the machine-specific behavior was investigated by processing an unreinforced and glass-fiber-reinforced polyamide. The results obtained provide crucial insights into machine-specific behavior, which may, for instance, account for disparities between computer fluid dynamic (CFD) simulations and experimental results. Furthermore, it is expected that the description of the machine-specific behavior can contribute to transfer knowledge when applying transfer learning algorithms. Looking ahead to future research, it is advised to create what is referred to as a "machine fingerprint", and this proposal is accompanied by some preliminary recommendations for its development.

Deep transfer learning for visual analysis and attribution of paintings by Raphael

Visual analysis and authentication of artworks are challenging tasks central to art history and criticism. This preliminary study presents a computational tool for scholars examining and authenticating a restricted class of paintings, with a specific focus on the paintings of Raffaello Sanzio da Urbino, more popularly known as Raphael. We applied transfer learning to the ResNet50 deep neural network for feature extraction and used a support vector machine (SVM) binary classifier in support of authentication. Edge detection and analysis algorithms, considered to be crucial for capturing the essence of Raphael’s artistic style, including the brushwork signatures, were also integrated and are used as an authentication tool. The machine learning approach we have developed demonstrates an accuracy of 98% in image-based classification tasks during validation using a test set of well known and authentic paintings by Raphael. Of course, a full authentication protocol relies on provenance, history, material studies, iconography, studies of a work’s condition, and more. Our work, then, contributes to just a portion of a full authentication protocol. Our findings suggest that machine learning methods, properly employed by experts aware of context, may enhance and expand traditional visual analysis for problems in art authentication.

Self-Supervised Domain-Adaptive learning for Self-Labeling unknown rice grains during actual rice transportation process

Industrial rice agriculture has transitioned towards intelligent computing, with an increasing likelihood of incorporating unmanned decision-making systems. However, one lingering issue that persists is the continued reliance on annotations for activity description. Applying transfer learning and domain-adaptive learning to non-labeled target data with the assistance of source data can be explicitly, but depends on mitigating domain discrepancy instead of principally generating predictive. This compromise both target robustness and comprehensibility. Our aim to enable a model to rely effectively on explicit source knowledge and self-labeling to reliably self-annotate. Our proposed self-supervised domain adaptation method provides debiasing by utilizing both former and present states, which reduces inconsistency when generating imbalanced pseudo-labels and implicitly utilizes source knowledge to consistently increase target-specific knowledge. Comprehensive results show improved performance by 7.5% for public data benchmarks and a gain of 2.4% for a rice transportation scenario. Ablation studies reveal improvements of up to 8.2% in Top-1 accuracy and 2.5% in Top-5 accuracy compared to an existing debiasing method, and highlights substantial parameter and FLOPs saving of up to 4.2 and 14 times, respectively. This underscores the feasibility of our system over self-label-dependent learning.

Expansion of the CT-scans image set based on the pretrained DCGAN for improving the performance of the CNN

Diagnosis of COVID-19 based on CT scans images using artificial intelligence can massively reduce medical manpower, time, and resources and reduce misdiagnosis rate. However, there are few relevant image sets publicly available at present, which is not conducive to the training of models such as classification. Transfer learning has been extensively considered in image classification, but less utilized in image generation. This work attempts to expand this dataset with discriminators using Deep Convolutional Generative Adversarial Network (DCGAN) models of Imagenet pre-training models and verify the significance of pre-training models for generating image veracity. A DCGAN model using the Imagenet pre-training model (i.e. resnet18) and a DCGAN network without the pre-train model generated 100 positive and 100 negative images to extend the original dataset, respectively. These images were classified using a Convolutional Neural Network (CNN) binary classifier, and the accuracy improved from 82.67% to 85.33% after adding the pre-training model. Furthermore, the comparison shown by the Gradcam visualization shows that the discriminator with the pre-training model can better capture the key details of the images. This experiment shows that even though Imagenet and lung images are totally uncorrelated, there are still some features that can be transferred. It also demonstrates that Imagenet pre-training models can improve the quality of GAN-generated images, extending the application of migration transfer learning.

AutoKE: An Automatic Knowledge Embedding Framework for Scientific Machine Learning

Imposing physical constraints on neural networks as a method of knowledge embedding has achieved great progress in solving physical problems described by governing equations. However, for many engineering problems, governing equations often have complex forms, including complex partial derivatives or stochastic physical fields, which results in significant inconveniences from the perspective of implementation. In this paper, to effectively automate the process of embedding physical knowledge, a scientific machine learning framework, called AutoKE, is proposed, and a reservoir flow problem with a complex governing equation is taken as an instance. In AutoKE, an emulator comprised of deep neural networks (DNNs) is built for predicting the physical variables of interest. An arbitrarily complex equation can be parsed and automatically converted into a computational graph through the equation parser module, and the fitness of the emulator to the governing equation is evaluated via automatic differentiation. Furthermore, the fixed weights in the loss function are substituted with adaptive weights by incorporating the Lagrangian dual method. Neural architecture search (NAS) is also introduced into the AutoKE to select an optimal network architecture of the emulator according to the specific problem. Finally, we apply transfer learning to enhance the scalability of the emulator. In experiments, the framework is verified by a series of physical problems in which it can automatically embed physical knowledge into an emulator without heavy hand-coding. The results demonstrate that the emulator can not only make accurate predictions, but also be applied to similar problems with high efficiency via transfer learning.

Prediction of Short-Shot Defects in Injection Molding by Transfer Learning

For a long time, the traditional injection molding industry has faced challenges in improving production efficiency and product quality. With advancements in Computer-Aided Engineering (CAE) technology, many factors that could lead to product defects have been eliminated, reducing the costs associated with trial runs during the manufacturing process. However, despite the progress made in CAE simulation results, there still exists a slight deviation from actual conditions. Therefore, relying solely on CAE simulations cannot entirely prevent product defects, and businesses still need to implement real-time quality checks during the production process. In this study, we developed a Back Propagation Neural Network (BPNN) model to predict the occurrence of short-shots defects in the injection molding process using various process states as inputs. We developed a Back Propagation Neural Network (BPNN) model that takes injection molding process states as input to predict the occurrence of short-shot defects during the injection molding process. Additionally, we investigated the effectiveness of two different transfer learning methods. The first method involved training the neural network model using CAE simulation data for products with length–thickness ratios (LT) of 60 and then applying transfer learning with real process data. The second method trained the neural network model using real process data for products with LT60 and then applied transfer learning with real process data from products with LT100. From the results, we have inferred that transfer learning, as compared to conventional neural network training methods, can prevent overfitting with the same amount of training data. The short-shot prediction models trained using transfer learning achieved accuracies of 90.2% and 94.4% on the validation datasets of products with LT60 and LT100, respectively. Through integration with the injection molding machine, this enables production personnel to determine whether a product will experience a short-shot before the mold is opened, thereby increasing troubleshooting time.

Application of transfer learning to predict drug-induced human in vivo gene expression changes using rat in vitro and in vivo data.

The liver is the primary site for the metabolism and detoxification of many compounds, including pharmaceuticals. Consequently, it is also the primary location for many adverse reactions. As the liver is not readily accessible for sampling in humans; rodent or cell line models are often used to evaluate potential toxic effects of a novel compound or candidate drug. However, relating the results of animal and in vitro studies to relevant clinical outcomes for the human in vivo situation still proves challenging. In this study, we incorporate principles of transfer learning within a deep artificial neural network allowing us to leverage the relative abundance of rat in vitro and in vivo exposure data from the Open TG-GATEs data set to train a model to predict the expected pattern of human in vivo gene expression following an exposure given measured human in vitro gene expression. We show that domain adaptation has been successfully achieved, with the rat and human in vitro data no longer being separable in the common latent space generated by the network. The network produces physiologically plausible predictions of human in vivo gene expression pattern following an exposure to a previously unseen compound. Moreover, we show the integration of the human in vitro data in the training of the domain adaptation network significantly improves the temporal accuracy of the predicted rat in vivo gene expression pattern following an exposure to a previously unseen compound. In this way, we demonstrate the improvements in prediction accuracy that can be achieved by combining data from distinct domains.

Deep learning and transfer learning of earthquake and quarry-blast discrimination: applications to southern California and eastern Kentucky

SUMMARY Discrimination between tectonic earthquakes (EQs) and quarry blasts is important for accurate EQ cataloguing and seismic hazard analysis. However, reliable classification of these two types of seismic events is challenging with no prior knowledge of source parameters. Here, we applied deep learning to perform this classification task in southern California and eastern Kentucky. Since the two regions differ significantly in available labelled data, class imbalance and waveform characteristics, we adopted different strategies for them. We directly trained a convolutional neural network (CNN) for southern California due to its data abundancy. To alleviate the class imbalance, the blast data were augmented by repeated sampling. The model for California yields F1-scores of >83.5 per cent when estimated by individual stations and >98.1 per cent by network average (i.e. averaging the CNN’s outputs on all available stations for each event). As eastern Kentucky has a much smaller data size, we apply transfer learning to the pre-trained California model to fit the Kentucky data. The transfer-learned model yields F1-scores of >86.9 per cent when estimated by individual stations and >96.7 per cent by network average. The transfer-learned model outperforms the model re-trained from scratch for the Kentucky data. Gradient-weighted class activation mapping shows the S onset and the S long-period coda are important to identify EQs and blasts, respectively. By visual inspections of the seismograms, the source locations, the origin time and the P-wave polarities, we verified that most of the events falsely predicted by our models are actually mislabelled by seismic analysts. Our models thus show great potential in helping seismic analysts find those mislabelled events which remain hidden in the historical catalogue. Our results demonstrate that deep learning can achieve high accuracy in seismic event discrimination and that transfer learning is effective and efficient to generalize deep learning models across different regions.

Deep Transfer Learning for Detecting Covid-19, Pneumonia and Tuberculosis Using Cxr Images – A Review

Chest X-rays remains to be the most common imaging modality used to diagnose lung diseases. However, they necessitate the interpretation of experts (radiologists and pulmonologists), who are few. This review paper investigates the use of deep transfer learning techniques to detect COVID-19, pneumonia, and tuberculosis in chest X-ray (CXR) images. It provides an overview of current state-of the-art CXR image classification techniques and discusses the challenges and opportunities in applying transfer learning to this domain. The paper provides a thorough examination of recent research studies that used deep transfer learning algorithms for COVID-19, pneumonia, and tuberculosis detection, highlighting the advantages and disadvantages of these approaches. Finally, the review paper discusses future research directions in the field of deep transfer learning for CXR image classification, as well as the potential for these techniques to aid in the diagnosis and treatment of lung diseases.