Medical Image Classification Research Articles

Semi-Supervised Medical Image Classification with Pseudo Labels Using Coalition Similarity Training

The development of medical image classification models necessitates a substantial number of labeled images for model training. In real-world scenarios, sample sizes are typically limited and labeled samples often constitute only a small portion of the dataset. This paper aims to investigate a collaborative similarity learning strategy that optimizes pseudo-labels to enhance model accuracy and expedite its convergence, known as the joint similarity learning framework. By integrating semantic similarity and instance similarity, the pseudo-labels are mutually refined to ensure their quality during initial training. Furthermore, the similarity score is utilized as a weight to guide samples away from misclassification predictions during the classification process. To enhance the model’s generalization ability, an adaptive consistency constraint is introduced into the loss function to improve performance on untrained datasets. The model achieved a satisfactory accuracy of 93.65% at 80% labeling ratio, comparable to supervised learning methods’ performance. Even with very low labeling ratio (e.g., 5%), the model still attained an accuracy of 74.28%. Comparison with other techniques such as Mean Teacher and FixMatch revealed that our approach significantly outperforms them in medical image classification tasks through improving accuracy by approximately 2%, demonstrating this framework’s leadership in medical image classification.

KDViT: COVID-19 diagnosis on CT-scans with knowledge distillation of vision transformer

This paper introduces Knowledge Distillation of Vision Transformer (KDViT), a novel approach for medical image classification. The Vision Transformer architecture incorporates a self-attention mechanism to autonomously learn image structure. The input medical image is segmented into patches and transformed into low-dimensional linear embeddings. Position information is integrated into each patch, and a learnable classification token is appended for classification, thereby preserving spatial relationships within the image. The output vectors are then fed into a Transformer encoder to extract both local and global features, leveraging the inherent attention mechanism for robust feature extraction across diverse medical imaging scenarios. Furthermore, knowledge distillation is employed to enhance performance by transferring insights from a large teacher model to a small student model. This approach reduces the computational requirements of the larger model and improves overall effectiveness. Integrating knowledge distillation with two Vision Transformer models not only showcases the novelty of the proposed solution for medical image classification but also enhances model interpretability, reduces computational complexity, and improves generalization capabilities. The proposed KDViT model achieved high accuracy rates of 98.39%, 88.57%, and 99.15% on the SARS-CoV-2-CT, COVID-CT, and iCTCF datasets respectively, surpassing the performance of other state-of-the-art methods.

Shifting to machine supervision: annotation-efficient semi and self-supervised learning for automatic medical image segmentation and classification

Advancements in clinical treatment are increasingly constrained by the limitations of supervised learning techniques, which depend heavily on large volumes of annotated data. The annotation process is not only costly but also demands substantial time from clinical specialists. Addressing this issue, we introduce the S4MI (Self-Supervision and Semi-Supervision for Medical Imaging) pipeline, a novel approach that leverages advancements in self-supervised and semi-supervised learning. These techniques engage in auxiliary tasks that do not require labeling, thus simplifying the scaling of machine supervision compared to fully-supervised methods. Our study benchmarks these techniques on three distinct medical imaging datasets to evaluate their effectiveness in classification and segmentation tasks. Notably, we observed that self-supervised learning significantly surpassed the performance of supervised methods in the classification of all evaluated datasets. Remarkably, the semi-supervised approach demonstrated superior outcomes in segmentation, outperforming fully-supervised methods while using 50% fewer labels across all datasets. In line with our commitment to contributing to the scientific community, we have made the S4MI code openly accessible, allowing for broader application and further development of these methods. The code can be accessed at https://github.com/pranavsinghps1/S4MI.

Automatic classification of spinal osteosarcoma and giant cell tumor of bone using optimized DenseNet

ObjectiveThis study aims to explore an optimized deep-learning model for automatically classifying spinal osteosarcoma and giant cell tumors. In particular, it aims to provide a reliable method for distinguishing between these challenging diagnoses in medical imaging. MethodsThis research employs an optimized DenseNet model with a self-attention mechanism to enhance feature extraction capabilities and reduce misclassification in differentiating spinal osteosarcoma and giant cell tumors. The model utilizes multi-scale feature map extraction for improved classification accuracy. The paper delves into the practical use of Gradient-weighted Class Activation Mapping (Grad-CAM) for enhancing medical image classification, specifically focusing on its application in diagnosing spinal osteosarcoma and giant cell tumors. The results demonstrate that the implementation of Grad-CAM visualization techniques has improved the performance of the deep learning model, resulting in an overall accuracy of 85.61%. Visualizations of images for these medical conditions using Grad-CAM, with corresponding class activation maps that indicate the tumor regions where the model focuses during predictions. ResultsThe model achieves an overall accuracy of 80% or higher, with sensitivity exceeding 80% and specificity surpassing 80%. The average area under the curve AUC for spinal osteosarcoma and giant cell tumors is 0.814 and 0.882, respectively. The model significantly supports orthopedics physicians in developing treatment and care plans. ConclusionThe DenseNet-based automatic classification model accurately distinguishes spinal osteosarcoma from giant cell tumors. This study contributes to medical image analysis, providing a valuable tool for clinicians in accurate diagnostic classification. Future efforts will focus on expanding the dataset and refining the algorithm to enhance the model's applicability in diverse clinical settings.

Boundary sample-based class-weighted semi-supervised learning for malignant tumor classification of medical imaging.

Medical image classification plays a pivotal role within the field of medicine. Existing models predominantly rely on supervised learning methods, which necessitate large volumes of labeled data for effective training. However, acquiring and annotating medical image data is both an expensive and time-consuming endeavor. In contrast, semi-supervised learning methods offer a promising approach by harnessing limited labeled data alongside abundant unlabeled data to enhance the performance of medical image classification. Nonetheless, current methods often encounter confirmation bias due to noise inherent in self-generated pseudo-labels and the presence of boundary samples from different classes. To overcome these challenges, this study introduces a novel framework known as boundary sample-based class-weighted semi-supervised learning (BSCSSL) for medical image classification. Our method aims to alleviate the impact of intra- and inter-class boundary samples derived from unlabeled data. Specifically, we address reliable confidential data and inter-class boundary samples separately through the utilization of an inter-class boundary sample mining module. Additionally, we implement an intra-class boundary sample weighting mechanism to extract class-aware features specific to intra-class boundary samples. Rather than discarding such intra-class boundary samples outright, our approach acknowledges their intrinsic value despite the difficulty associated with accurate classification, as they contribute significantly to model prediction. Experimental results on widely recognized medical image datasets demonstrate the superiority of our proposed BSCSSL method over existing semi-supervised learning approaches. By enhancing the accuracy and robustness of medical image classification, our BSCSSL approach yields considerable implications for advancing medical diagnosis and future research endeavors.

Preoperative Molecular Subtype Classification Prediction of Ovarian Cancer Based on Multi-Parametric Magnetic Resonance Imaging Multi-Sequence Feature Fusion Network.

In the study of the deep learning classification of medical images, deep learning models are applied to analyze images, aiming to achieve the goals of assisting diagnosis and preoperative assessment. Currently, most research classifies and predicts normal and cancer cells by inputting single-parameter images into trained models. However, for ovarian cancer (OC), identifying its different subtypes is crucial for predicting disease prognosis. In particular, the need to distinguish high-grade serous carcinoma from clear cell carcinoma preoperatively through non-invasive means has not been fully addressed. This study proposes a deep learning (DL) method based on the fusion of multi-parametric magnetic resonance imaging (mpMRI) data, aimed at improving the accuracy of preoperative ovarian cancer subtype classification. By constructing a new deep learning network architecture that integrates various sequence features, this architecture achieves the high-precision prediction of the typing of high-grade serous carcinoma and clear cell carcinoma, achieving an AUC of 91.62% and an AP of 95.13% in the classification of ovarian cancer subtypes.

Suppressing label noise in medical image classification using mixup attention and self-supervised learning

Deep neural networks (DNNs) have been widely applied in medical image classification and achieve remarkable classification performance. These achievements heavily depend on large-scale accurately annotated training data. However, label noise is inevitably introduced in the medical image annotation, as the labeling process heavily relies on the expertise and experience of annotators. Meanwhile, DNNs suffer from overfitting noisy labels, degrading the performance of models. Therefore, in this work, we innovatively devise a noise-robust training approach to mitigate the adverse effects of noisy labels in medical image classification. Specifically, we incorporate contrastive learning and intra-group mixup attention strategies into vanilla supervised learning. The contrastive learning for feature extractor helps to enhance visual representation of DNNs. The intra-group mixup attention module constructs groups and assigns self-attention weights for group-wise samples, and subsequently interpolates massive noisy-suppressed samples through weighted mixup operation. We conduct comparative experiments on both synthetic and real-world noisy medical datasets under various noise levels. Rigorous experiments validate that our noise-robust method with contrastive learning and mixup attention can effectively handle with label noise, and is superior to state-of-the-art methods. An ablation study also shows that both components contribute to boost model performance. The proposed method demonstrates its capability of curb label noise and has certain potential toward real-world clinic applications.

MTUNet + + : explainable few-shot medical image classification with generative adversarial network

MTUNet + + : explainable few-shot medical image classification with generative adversarial network

Enhancing medical image classification with generative AI using latent denoising diffusion probabilistic model and wiener filtering approach

In the evolving landscape of medical diagnostics, a paradigm shift is being catalysed by the advent of generative artificial intelligence. Medical X-ray, CT and MRI images are essential diagnostic tools used by healthcare professionals to assess various musculoskeletal conditions. However, obtaining a sufficient number of medical images for training deep learning models can be challenging due to limited access to labelled data. Hence, the authors propose a novel Latent diffusion process for synthesizing medical images that closely resemble real patient images, aiming to address the challenge of limited access to labelled data in medical diagnostics. Leveraging deep learning and generative modelling techniques, this method synthesizes high-fidelity images that closely mimic real patient scans. By introducing noise and subsequently training the model to denoise, the approach captures intricate patterns inherent in authentic medical images. Among the four datasets utilized, the Diabetes Retinopathy dataset demonstrates superior performance, achieving the highest Mean Structural Similarity Index (MSSIM) score of 0.57 (compared to the dataset baseline of 0.62) and an accuracy of 93.75% when passed through the proposed pipeline. The Cataract dataset, registered a MSSIM score of 0.51 (versus the dataset baseline of 0.53) and an accuracy score of 97.52%, while the Knee OA dataset follows closely with MSSIM and accuracy scores of 0.65 (in contrast to the dataset baseline of 0.63) and 68.66% respectively. The results obtained are then compared with the results generated by the other state of the art models.

A prior knowledge-guided distributionally robust optimization-based adversarial training strategy for medical image classification

Medical image classification plays a vital role in computer vision applications within the field of healthcare and medicine, and deep neural network-based classifiers are continuously achieving new breakthroughs and demonstrating tremendous potential in medical imaging analysis. However, the lack of robustness in deep learning techniques makes it risky to apply these classifiers to the domain of healthcare. In addition, the existing adversarial training strategies and domain generation methods are difficult to generalize into the medical imaging field challenged by complex medical texture features. To address this issue, we propose a medical morphological knowledge-guided adversarial training strategy, by jointly considering the robustness against medical data distribution shifts and adversarial attacks from the view of distributionally robust optimization. First, we train a surrogate model with the augmented dataset by guided filtering for capturing model attention on medical morphological information. Next, we design a gradient normalization-based prior knowledge injection module to transfer the attention information learned by surrogate model to the main classifier. Finally, we design a distributionally robust a optimization-based training strategy to induce the main classifier to learn key diagnostic clues as well as enhance the robustness against adversarial attacks. To evaluate the effectiveness of the proposed methods, we perform experiments on two types of in-domain and out-of-domain medical image sets, which contain lung CT scan datasets and dermatoscopic image datasets. Comparative results show that the proposed training strategy achieves higher adversarial attack accuracy than all involved state-of-the-art adversarial training methods and domain generation methods. The code is available at https://github.com/sysu19351146/MMK-DRO.

A comparative study of an on premise AutoML solution for medical image classification

Automated machine learning (AutoML) allows for the simplified application of machine learning to real-world problems, by the implicit handling of necessary steps such as data pre-processing, feature engineering, model selection and hyperparameter optimization. This has encouraged its use in medical applications such as imaging. However, the impact of common parameter choices such as the number of trials allowed, and the resolution of the input images, has not been comprehensively explored in existing literature. We therefore benchmark AutoKeras (AK), an open-source AutoML framework, against several bespoke deep learning architectures, on five public medical datasets representing a wide range of imaging modalities. It was found that AK could outperform the bespoke models in general, although at the cost of increased training time. Moreover, our experiments suggest that a large number of trials and higher resolutions may not be necessary for optimal performance to be achieved.

Typicality- and instance-dependent label noise-combating: a novel framework for simulating and combating real-world noisy labels for endoscopic polyp classification

Learning with noisy labels aims to train neural networks with noisy labels. Current models handle instance-independent label noise (IIN) well; however, they fall short with real-world noise. In medical image classification, atypical samples frequently receive incorrect labels, rendering instance-dependent label noise (IDN) an accurate representation of real-world scenarios. However, the current IDN approaches fail to consider the typicality of samples, which hampers their ability to address real-world label noise effectively. To alleviate the issues, we introduce typicality- and instance-dependent label noise (TIDN) to simulate real-world noise and establish a TIDN-combating framework to combat label noise. Specifically, we use the sample’s distance to decision boundaries in the feature space to represent typicality. The TIDN is then generated according to typicality. We establish a TIDN-attention module to combat label noise and learn the transition matrix from latent ground truth to the observed noisy labels. A recursive algorithm that enables the network to make correct predictions with corrections from the learned transition matrix is proposed. Our experiments demonstrate that the TIDN simulates real-world noise more closely than the existing IIN and IDN. Furthermore, the TIDN-combating framework demonstrates superior classification performance when training with simulated TIDN and actual real-world noise.

Enhancing Open-World Bacterial Raman Spectra Identification by Feature Regularization for Improved Resilience against Unknown Classes.

The combination of deep learning techniques and Raman spectroscopy shows great potential offering precise and prompt identification of pathogenic bacteria in clinical settings. However, the traditional closed-set classification approaches assume that all test samples belong to one of the known pathogens, and their applicability is limited since the clinical environment is inherently unpredictable and dynamic, unknown, or emerging pathogens may not be included in the available catalogs. We demonstrate that the current state-of-the-art neural networks identifying pathogens through Raman spectra are vulnerable to unknown inputs, resulting in an uncontrollable false positive rate. To address this issue, first we developed an ensemble of ResNet architectures combined with the attention mechanism that achieves a 30-isolate accuracy of 87.8 ± 0.1%. Second, through the integration of feature regularization by the Objectosphere loss function, our model both achieves high accuracy in identifying known pathogens from the catalog and effectively separates unknown samples drastically reducing the false positive rate. Finally, the proposed feature regularization method during training significantly enhances the performance of out-of-distribution detectors during the inference phase improving the reliability of the detection of unknown classes. Our algorithm for Raman spectroscopy empowers the identification of previously unknown, uncataloged, and emerging pathogens ensuring adaptability to future pathogens that may surface. Moreover, it can be extended to enhance open-set medical image classification, bolstering its reliability in dynamic operational settings.

Acute lymphocytic leukemia detection and subtype classification via extended wavelet pooling based-CNNs and statistical-texture features

Acute lymphoblastic leukemia (ALL) is considered the most fatal form of leukemia (also known as blood cancer). It propagates quickly among adults and children and could lead to their death. Early detection of ALL and ALL subtypes is the key factor in selecting effective treatment types and improving survival rates. However, routine diagnostic approaches have several drawbacks. Computer-assisted diagnosis (CAD) is the perfect solution to avoid these challenges and achieve a fast and accurate diagnosis. Current CAD models require enhancement/segmentation processing. Besides, they are either dependent on deep learning (DL) models or handcrafted features along with machine learning. Those CADs that employed DL approaches relied solely on spatial information during the training procedure. However, learning them with spectral temporal and temporal representations could improve performance. Furthermore, integrating deep features from DL models along with handcrafted features can increase the discrimination ability of attributes in medical image classification. This study aims to propose a novel CAD for ALL detection and subtype classification without pre-segmentation or enhancement steps. The proposed CAD extends the conventional DL models of convolutional neural networks by introducing an additional wavelet pooling, accompanied by a dense layer or a long-short-term memory (LSTM) layer, and then a SoftMax layer, acquiring spectral-temporal information along with temporal information. To further improve the framework's ability to discriminate, the introduced CAD then combines the wavelet-based deep features of every CNN with numerous handcrafted attributes. Afterward, a feature selection methodology is utilized to create a model with limited features and improved accuracy. The performance results show that the novel CAD is capable of achieving 100% ALL detection accuracy, as well as 100% ALL-subtype classification accuracy with just 88 and 146 features. Thus, this CAD can be employed to assist pathologists in the rapid and precise ALL identification and subcategories recognition.

Medical image Classification using Transfer Learning: Convolutional Neural Network Models approach

A relatively low life expectancy in their greatest grade is caused by more debilitating diseases in the medical sector. Any kind of misdiagnosis may lead to improper medical intervention and lower the likelihood that a patient will survive. Making an appropriate treatment plan to cure and enhance the quality of life for patients suffering from any kind of illness starts with an accurate diagnosis. Convolutional neural networks and computer-aided disease detection systems have produced success stories and advanced the science of machine learning significantly. Compared to more conventional neural network layers, the deep convolutional layers automatically extract significant and reliable characteristics from the input space. Within the suggested structure, we use two convolutional neural network architectures (VGG16 and VGG19) to perform two types of medical imaging (x-ray and MRI) investigations in order to categorize brain tumors as normal and up normal as well as classify x-ray pneumonia or normal. The transfer learning strategies by VGG16 and vgg19 models that is, using MRI and x-ray to fine-tune and freeze are then examined in each investigation. In order to increase dataset samples and decrease the likelihood of over-fitting, data augmentation techniques are performed to MRI slices and x-rays for results that can be more broadly utilized. In the proposed studies, the fine-tune VGG19, VGG16 architecture attained highest accuracy up to 0.95with x-ray and 0.80 respectively. The accuracy0.98 used MRI in terms of classification and detection with VGG19.

Dynamic Weighting Translation Transfer Learning for Imbalanced Medical Image Classification.

Medical image diagnosis using deep learning has shown significant promise in clinical medicine. However, it often encounters two major difficulties in real-world applications: (1) domain shift, which invalidates the trained model on new datasets, and (2) class imbalance problems leading to model biases towards majority classes. To address these challenges, this paper proposes a transfer learning solution, named Dynamic Weighting Translation Transfer Learning (DTTL), for imbalanced medical image classification. The approach is grounded in information and entropy theory and comprises three modules: Cross-domain Discriminability Adaptation (CDA), Dynamic Domain Translation (DDT), and Balanced Target Learning (BTL). CDA connects discriminative feature learning between source and target domains using a synthetic discriminability loss and a domain-invariant feature learning loss. The DDT unit develops a dynamic translation process for imbalanced classes between two domains, utilizing a confidence-based selection approach to select the most useful synthesized images to create a pseudo-labeled balanced target domain. Finally, the BTL unit performs supervised learning on the reassembled target set to obtain the final diagnostic model. This paper delves into maximizing the entropy of class distributions, while simultaneously minimizing the cross-entropy between the source and target domains to reduce domain discrepancies. By incorporating entropy concepts into our framework, our method not only significantly enhances medical image classification in practical settings but also innovates the application of entropy and information theory within deep learning and medical image processing realms. Extensive experiments demonstrate that DTTL achieves the best performance compared to existing state-of-the-art methods for imbalanced medical image classification tasks.

Disentangle Then Calibrate With Gradient Guidance: A Unified Framework for Common and Rare Disease Diagnosis.

The computer-aided diagnosis (CAD) for rare diseases using medical imaging poses a significant challenge due to the requirement of large volumes of labeled training data, which is particularly difficult to collect for rare diseases. Although Few-shot learning (FSL) methods have been developed for this task, these methods focus solely on rare disease diagnosis, failing to preserve the performance in common disease diagnosis. To address this issue, we propose the Disentangle then Calibrate with Gradient Guidance (DCGG) framework under the setting of generalized few-shot learning, i.e., using one model to diagnose both common and rare diseases. The DCGG framework consists of a network backbone, a gradient-guided network disentanglement (GND) module, and a gradient-induced feature calibration (GFC) module. The GND module disentangles the network into a disease-shared component and a disease-specific component based on gradient guidance, and devises independent optimization strategies for both components, respectively, when learning from rare diseases. The GFC module transfers only the disease-shared channels of common-disease features to rare diseases, and incorporates the optimal transport theory to identify the best transport scheme based on the semantic relationship among different diseases. Based on the best transport scheme, the GFC module calibrates the distribution of rare-disease features at the disease-shared channels, deriving more informative rare-disease features for better diagnosis. The proposed DCGG framework has been evaluated on three public medical image classification datasets. Our results suggest that the DCGG framework achieves state-of-the-art performance in diagnosing both common and rare diseases.

DeepChestGNN: A Comprehensive Framework for Enhanced Lung Disease Identification through Advanced Graphical Deep Features.

Lung diseases are the third-leading cause of mortality in the world. Due to compromised lung function, respiratory difficulties, and physiological complications, lung disease brought on by toxic substances, pollution, infections, or smoking results in millions of deaths every year. Chest X-ray images pose a challenge for classification due to their visual similarity, leading to confusion among radiologists. To imitate those issues, we created an automated system with a large data hub that contains 17 datasets of chest X-ray images for a total of 71,096, and we aim to classify ten different disease classes. For combining various resources, our large datasets contain noise and annotations, class imbalances, data redundancy, etc. We conducted several image pre-processing techniques to eliminate noise and artifacts from images, such as resizing, de-annotation, CLAHE, and filtering. The elastic deformation augmentation technique also generates a balanced dataset. Then, we developed DeepChestGNN, a novel medical image classification model utilizing a deep convolutional neural network (DCNN) to extract 100 significant deep features indicative of various lung diseases. This model, incorporating Batch Normalization, MaxPooling, and Dropout layers, achieved a remarkable 99.74% accuracy in extensive trials. By combining graph neural networks (GNNs) with feedforward layers, the architecture is very flexible when it comes to working with graph data for accurate lung disease classification. This study highlights the significant impact of combining advanced research with clinical application potential in diagnosing lung diseases, providing an optimal framework for precise and efficient disease identification and classification.

A Study of Neutrosophic Sets and Deep Learning Models for Breast Cancer Classification

Medical image classification and detection using artificial intelligence (AI) can help enhance medical care services. Unfortunately, most medical image modalities suffer from some noise and uncertainty, which decreases the performance of disease detection and classification. Neutrosophic sets (NS) can handle uncertainty data within medical images. NS can present images into three subsets: true (T), indeterminacy (I), and fuzzy sets (FS). In this study, we investigate the performance of deep learning (DL) models under the NS domain for breast cancer classification. The crisp image domain is converted to the NS domain and represents the image into three subsets. The converted images are used to train seven DL models, such as VGG16, VGG19, ResNet50, ResNet150, DenseNest121, EfficientNetB2, and MobileNetV2. A comparison between DL under NS and on the FS, domain has been made in terms of accuracy, precision, recall, and F1 score. The experimental results showed superior results for NS over FS.

A Novel Hybrid Approach: Capsule Network Enhanced with Residual Block and CBAM Module for Medical Image Classification

A Novel Hybrid Approach: Capsule Network Enhanced with Residual Block and CBAM Module for Medical Image Classification