Classification Of Microscopic Images Research Articles

Should We Believe in Cells, or Just Remain Agnostic about Them? A Critical Analysis Through Bas van Fraassen’s Lenses.

For constructive empiricism, being observable or unobservable is defining for deciding about the empirical adequacy and epistemic value of theory components. The classification of microscopic images has been particularly debated. van Fraassen initially classified microscope images as unobservable and then as mere images, like rainbows. Afterwards, he claimed it is not irrational to maintain neutrality about their classification and left them in some kind of limbo between being images of something real or being mere images. Here, I provide an argument to classify microscopic images of cells as copy-qualified images. The argument is described in general terms to sustain that any unobservable entity that is derived from observable self-dividing entities and looks similar to the observable entity when detected under the microscope, corresponds to a copy-qualified image of something real. Because of the existence in nature of large observable cells that fulfil these properties, I conclude that microscopic images of cells are images of something real. This determines their empirical adequacy and epistemic value, making them hard-core and stable components of biological theories. The argument described provides a strategy to develop a more grained classification of what is observable and unobservable with the consequent implications for theory development.

Applying feature-similarity-metrics for long-tailed problem of phytoplankton microscopic images classification

The distribution of phytoplankton in natural water bodies holds significant importance for aquatic for maintaining healthy aquatic environments. While deep learning has become a popular research field of automating phytoplankton identification, its performance is diminished by the uneven species distribution in nature world, which biases classification models towards more advantaged species. Addressing this long-tailed problem, the paper proposed a novel deep learning idea termed feature-similarity-metrics species which utilizes the principle of decreasing intra-class differences and increasing the inter-class differences. A method called feature-center-constraint is introduced to implement the idea. Initially, a center is presupposed for each category, and these centers are distributed evenly and sparsely within the feature space using a uniform distribution initialization method. During model training, the Euclidean distance is used to measure the similarity between sample features and corresponding center. This approach enhanced the model's ability to predict disadvantaged phytoplankton species without compromising the classification accuracy of advantaged ones. Using a phytoplankton dataset of 26,564 images from 68 genera, collected from Chaohu Lake, this method demonstrates superior performance compared to existing techniques. Specially, it achieves a 5.69 % increase in micro-average F1 score and an 11.85 % rise in macro-average F1 score over general method, while F1 score is widely recognized as a reliable metric for measuring classification models' performance. These enhancements not only increase the accuracy of automated phytoplankton identification but also facilitate more effective monitoring of biodiversity and ecological health in aquatic systems. This makes the research highly beneficial for practical applications in environmental science, offering rapid and accurate insights into phytoplankton community structures, for ecological assessments.

Classification of Microscopic Hyperspectral Images of Blood Cells Based on Lightweight Convolutional Neural Network

Hyperspectral imaging has emerged as a novel imaging modality in the medical field, offering the ability to acquire images of biological tissues while simultaneously providing biochemical insights for in-depth tissue analysis. This approach facilitates early disease diagnosis, presenting advantages over traditional medical imaging techniques. Addressing challenges such as the computational burden of existing convolutional neural networks (CNNs) and imbalances in sample data, this paper introduces a lightweight GhostMRNet for the classification of microscopic hyperspectral images of human blood cells. The proposed model employs Ghost Modules to replace conventional convolutional layers and a cascading approach with small convolutional kernels for multiscale feature extraction, aiming to enhance feature extraction capabilities while reducing computational complexity. Additionally, an SE (Squeeze-and-Excitation) module is introduced to selectively allocate weights to features in each channel, emphasizing informative features and efficiently achieving spatial–spectral feature extraction in microscopic hyperspectral imaging. We evaluated the performance of the proposed GhostMRNet and compared it with other state-of-the-art models using two real medical hyperspectral image datasets. The experimental results demonstrate that GhostMRNet exhibits a superior performance, with an overall accuracy (OA), average accuracy (AA), and Kappa coefficient reaching 99.965%, 99.565%, and 0.9925, respectively. In conclusion, the proposed GhostMRNet achieves a superior classification performance at a smaller computational cost, thereby providing a novel approach for blood cell detection.

Recognition and classification of microscopic fatigue fracture images of high-strength bolt using deep learning methods

The fracture surface of high-strength bolt after fatigue fracture contains a lot of information, such as the location of stress concentration and the distribution of fatigue cracks. In this study, a large number of scanning electron microscope (SEM) images of fatigue fracture surface of broken high-strength bolt were identified and classified using the method of deep learning. At the beginning, a data set of SEM images containing 1556 fatigue fractures of high-strength bolts was prepared. Then, three convolutional neural networks, VGG16, ResNets50 and MobileNets, were used to recognize and classify the images in the dataset. In this process, part of the convolution layer of ResNets50 was extracted for visualization. At the same time, the Loss-Epoch curves, accuracy, recall and confusion matrices of the three networks were derived to evaluate the nets. Finally, the network with the highest accuracy was selected to adjust the parameters to further improve the accuracy of the classification. It was found that the three nets can complete the classification of these images. MobileNets had the best performance for this classification task, and the accuracy rate after adjusting the parameters has reached 86.76%. For some images with obvious features, the recall rate of classification had reached 100%. However, images from the same fatigue area were prone to a small amount of confusion. Finally, the feature map of the network would become more abstract with the deepening of the network, and the features of the image concerned by each convolution layer were also different.

Classification of Microscopic Fungi Images Using Vision Transformers for Enhanced Detection of Fungal Infections

Fungi play a pivotal role in our ecosystem and human health, serving as both essential contributors to environmental sustainability and significant agents of disease. The importance of precise fungi detection cannot be overstated, as it underpins effective disease management, agricultural productivity, and the safeguarding of global food security. This research explores the efficacy of vision transformer-based architectures for the classification of microscopic fungi images of various fungal types to enhance the detection of fungal infections. The study compared the pre-trained base Vision Transformer (ViT) and Swin Transformer models, evaluating their capability in feature extraction and fine-tuning. The incorporation of transfer learning and fine-tuning strategies, particularly with data augmentation, significantly enhances model performance. Utilizing a comprehensive dataset with and without data augmentation, the study reveals that Swin Transformer, particularly when fine-tuned, exhibits superior accuracy (98.36%) over ViT model (96.55%). These findings highlight the potential of vision transformer-based models in automating and refining the diagnosis of fungal infections, promising significant advancements in medical imaging analysis.

Efficient leukocytes detection and classification in microscopic blood images using convolutional neural network coupled with a dual attention network

Leukocytes, also called White Blood Cells (WBCs) or leucocytes, are the cells that play a pivotal role in human health and are vital indicators of diseases such as malaria, leukemia, AIDS, and other viral infections. WBCs detection and classification in blood smears offers insights to pathologists, aiding diagnosis across medical conditions. Traditional techniques, including manual counting, detection, classification, and visual inspection of microscopic images by medical professionals, pose challenges due to their labor-intensive nature. However, traditional methods are time consuming and sometimes susceptible to errors. Here, we propose a high-performance convolutional neural network (CNN) coupled with a dual-attention network that efficiently detects and classifies WBCs in microscopic thick smear images. The main aim of this study was to enhance clinical hematology systems and expedite medical diagnostic processes. In the proposed technique, we utilized a deep convolutional generative adversarial network (DCGAN) to overcome the limitations imposed by limited training data and employed a dual attention mechanism to improve accuracy, efficiency, and generalization. The proposed technique achieved overall accuracy rates of 99.83%, 99.35%, and 99.60% for the peripheral blood cell (PBC), leukocyte images for segmentation and classification (LISC), and Raabin-WBC benchmark datasets, respectively. Our proposed approach outperforms state-of-the-art methods in terms of accuracy, highlighting the effectiveness of the strategies employed and their potential to enhance diagnostic capabilities and advance real-world healthcare practices and diagnostic systems.

Enhanced label constrained contrastive learning for chromosome optical microscopic image classification

Karyotyping analysis, an indispensable means of diagnosing genetic diseases in clinical practice, entails a highly intricate process of chromosome classification. Automatic chromosome classification, in particular, poses a daunting challenge, impeding the acceleration of the diagnosis speed of karyotyping analysis. Hence, we propose a newfangled method of enhanced label-constrained contrastive learning to sensitively discriminate chromosomes. Our method incorporates a multi-scale feature perception aggregation encoder and a linear classifier and operates in two stages. The first stage involves the utilization of an encoder and projection head for chromosome representation learning. We optimally exploit label information and conduct contrastive learning on positive samples of the same category of chromosomes. Further, we fortify local irregularities by means of CutMix image augmentation and extract global and local features via multi-scale feature modules. The input perception aggregation module establishes feature dependencies and introduces a supervised loss to minimize intra-class distance and enlarge inter-class distance in the representation space. Finally, we map the learned representation to the Multi-Layer Perception (MLP) projection head for chromosome representation learning. In the second stage, we freeze the encoder, discard the projection head, and fine-tune the chromosome classification network by adding a classification layer that assigns each chromosome to a type in each patient case. We evaluate our method thoroughly on two public datasets and a private dataset, and the detailed results indicate that our proposed enhanced label-constrained contrastive learning attains the highest average accuracy of 99.272%. Our method outperforms state-of-the-art approaches, manifesting the effectiveness of our enhanced label-supervised contrastive learning, multi-scale feature integration, and feature dependency relationships. Additionally, it is also amenable to other medical image-related tasks.

Utilizing Autoencoders for Analysis and Classification of Microscopic in Situ Hybridization Images

Currently, analysis of microscopic In Situ Hybridization (ISH) images is done manually by experts. Precise evaluation and classification of such microscopic images can ease experts' work and reveal further insights about the data. In this work, we propose a deep-learning workflow to detect and classify areas of microscopic images with similar levels of gene expression. Analysis of the data is done by employing a type of ANN – Deep Learning Autoencoders – suitable for unsupervised learning. The model's performance is optimised by balancing the latent layers' length and complexity and fine-tuning hyperparameters. The results are validated by adapting the mean-squared error (MSE) metric and comparison to expert's evaluation. Reconstruction of the whole-scale microscopic images is used to summarise and visualise the results.

Classification of microscopic peripheral blood cell images using multibranch lightweight CNN-based model

Classification of microscopic peripheral blood cell images using multibranch lightweight CNN-based model

A robust classification of acute lymphocytic leukemia-based microscopic images with supervised Hilbert-Huang transform.

Acute lymphocytic leukemia (ALL) is a malignant condition characterized by the development of blast cells in the bone marrow and their quick dissemination into the bloodstream. It primarily affects children and individuals over the age of 60. Manual blood testing, which has been around for a long time, may be slow. The likelihood of recognizing ALL in its early stages was increased by automating the diagnosis. This research developed an improved criterion for classifying ALL microscopic images into two categories: normal images and blast images. First, to save processing time, innovative image preprocessing techniques were employed to gather data for data augmentation, enhancement, and conversion. The K-means clustering technique was also utilized to effectively segment the relevant nuclei from the background. Furthermore, the most salient features were extracted using an empirical mode decomposition (EMD) based on the Hilbert-Huang transform. MATLAB functions such as principal component analysis, gray level co-occurrence matrix, local binary pattern, shape features, discrete cosine transform, discrete Fourier transform, discrete wavelet transform, and independent component analysis have been used and compared with EMD. The Bayesian regularization (BR) method has been implemented in the neural networks (NNs) classifier. Along with NNs, other classifiers such as support vector machine, K-nearest neighbors, random forest, naive Bayes, logistic regression, and decision tree have been used, evaluated, and contrasted with NNs. According to experimental findings, the ALL-IDB2 (Image Database 2) dataset's NNs-based-EMD model classified objects with an accuracy of 98.7%, sensitivity of 99.3%, and specificity of 98.1%. RESEARCH HIGHLIGHTS: Implement a robust method for classifying normal and blast ALL images in the state of the art using the combination of the BR algorithm and the neural networks classifier. Perform robust data processing via data augmentation and conversion from RGB (Red, Green, and Blue) image LAB (Luminosity, A: color space, B: color space) image. Extract the nuclei correctly from the background image using k-means clustering. Extract the most salient features from the segmented images using EMD in the state of the art of HHT.

Microscopic Hyperspectral Image Classification Based on Fusion Transformer with Parallel CNN.

Microscopic hyperspectral image (MHSI) has received considerable attention in the medical field. The wealthy spectral information provides potentially powerful identification ability when combining with advanced convolutional neural network (CNN). However, for high-dimensional MHSI, the local connection of CNN makes it difficult to extract the long-range dependencies of spectral bands. Transformer overcomes this problem well because of its self-attention mechanism. Nevertheless, transformer is inferior to CNN in extracting spatial detailed features. Therefore, a classification framework integrating transformer and CNN in parallel, named as Fusion Transformer (FUST), is proposed for MHSI classification tasks. Specifically, the transformer branch is employed to extract the overall semantics and capture the long-range dependencies of spectral bands to highlight the key spectral information. The parallel CNN branch is designed to extract significant multiscale spatial features. Furthermore, the feature fusion module is developed to effectively fuse and process the features extracted by the two branches. Experimental results on three MHSI datasets demonstrate that the proposed FUST achieves superior performance when compared with state-of-the-art methods.

Blood Cell Microscopic Image Classification in Computer Aided Diagnosis Using Machine Learning: A Review

Blood cell detection considers a gold standard key in diagnosing blood disease and producing automatic reports to hematologists and doctors. Blood cell detection is a challenging task due to non-illumination level, high number of overlapped cells per image, variations in cell densities among platelets, white blood cells and red blood cells, and the variety of staining process. Traditional procedure of blood cell detection requires pathologist effort and time. In computer aided diagnosis, machine learning and deep learning techniques become the practical way to automate the procedure of diagnosing, classify microscopic blood cells, and increase the accuracy and speed of the procedure. This paper provides a review of the detection and classification of blood cell, including red blood cells, white blood cells and platelets and their characteristics using machine learning techniques. We also have detailed the dataset of microscope blood cell. We have divided the previous works into four categories based on the output of the models, including pre-processing, segmentation, feature extraction and classification. Then, we discuss the challenges that face these methods and suggest the potential future techniques.

Automatic Classification of Liquid Crystal Images Based on Topological Analysis

Liquid crystal image classification is a vital part of modern biosensors. Recent studies have shown that this problem can be solved using computationally intensive intelligent analysis tools, such as convolutional neural networks (CNNs) and support vector machines (SVMs). This article proposes a new method of microscopic image classification for liquid crystals-based biosensors with fast response. This method is based on topological analysis and provides 95% accuracy. Moreover, on the same hardware, it reaches eightfold performance compared to CNNs, which are usually used in similar applications. Finally, it has only nine parameters. Most of those parameters are independent and can be easily tuned based on the properties of the liquid crystals suspension and the microscope. This is a significant benefit compared to machine learning approaches that require large training datasets. The proposed solution can be considered a new step toward the creation of fully automatic biosensors for industrial water quality assessment systems.

Using image processing and automated classification models to classify microscopic gram stain images

Background and ObjectiveFast and correct classification of bacterial samples are important for accurate diagnostics and treatment. Manual microscopic interpretation of Gram stain samples is both time consuming and operator dependent. The aim of this study was to investigate the potential for developing an automated algorithm for the classification of microscopic Gram stain images. MethodsWe developed and tested two algorithms (using image processing an Casual Probabilistic Network (CPN) and a Random Forest (RF) classification) for the automated classification of Gram stain images. A dataset of 660 images including 33 microbial species (32 bacteria and one fungus) was split into training, validation, and test sets. The algorithms were evaluated based on their ability to correctly classify samples and general characteristics such as aggregation and morphology. ResultsThe CPN correctly classified 633/792 images to achieve an overall accuracy of 80% compared to the RF which correctly classified 782/792 images to achieve an overall accuracy of 99% (p < 0.001). The CPN performed well when distinguishing between GN and GP, with an accuracy of 95% (731/768). The RF also performed well in distinguishing between GN and GP, achieving an accuracy of 99% (767/768) (p < 0.001). ConclusionsThe findings from this study show promising results regarding the potential for an automated algorithm for the classification of microscopic Gram stain images.

A reliable and low-cost deep learning model integrating convolutional neural network and transformer structure for fine-grained classification of chicken Eimeria species

Chicken coccidiosis is a disease caused by Eimeria spp. and costs the broiler industry more than 14 billion dollars per year globally. Different chicken Eimeria species vary significantly in pathogenicity and virulence, so the classification of different chicken Eimeria species is of great significance for the epidemiological survey and related prevention and control. The microscopic morphological examination for their classification was widely used in clinical applications, but it is a time-consuming task and needs expertise. To increase the classification efficiency and accuracy, a novel model integrating transformer and convolutional neural network (CNN), named Residual-Transformer-Fine-Grained (ResTFG), was proposed and evaluated for fine-grained classification of microscopic images of seven chicken Eimeria species. The results showed that ResTFG achieved the best performance with high accuracy and low cost compared with traditional models. Specifically, the parameters, inference speed and overall accuracy of ResTFG are 1.95M, 256 FPS and 96.9%, respectively, which are 10.9 times lighter, 1.5 times faster and 2.7% higher in accuracy than the benchmark model. In addition, ResTFG showed better performance on the classification of the more virulent species. The results of ablation experiments showed that CNN or Transformer alone had model accuracies of only 89.8% and 87.0%, which proved that the improved performance of ResTFG was benefit from the complementary effect of CNN's local feature extraction and transformer's global receptive field. This study invented a reliable, low-cost, and promising deep learning model for the automatic fine-grain classification of chicken Eimeria species, which could potentially be embedded in microscopic devices to improve the work efficiency of researchers and extended to other parasite ova, and applied to other agricultural tasks as a backbone.

Advances Towards Automatic Detection and Classification of Parasites Microscopic Images Using Deep Convolutional Neural Network: Methods, Models and Research Directions.

In the developing world, parasites are responsible for causing several serious health problems, with relatively high infections in human beings. The traditional manual light microscopy process of parasite recognition remains the golden standard approach for the diagnosis of parasitic species, but this approach is time-consuming, highly tedious, and also difficult to maintain consistency but essential in parasitological classification for carrying out several experimental observations. Therefore, it is meaningful to apply deep learning to address these challenges. Convolution Neural Network and digital slide scanning show promising results that can revolutionize the clinical parasitology laboratory by automating the process of classification and detection of parasites. Image analysis using deep learning methods have the potential to achieve high efficiency and accuracy. For this review, we have conducted a thorough investigation in the field of image detection and classification of various parasites based on deep learning. Online databases and digital libraries such as ACM, IEEE, ScienceDirect, Springer, and Wiley Online Library were searched to identify sufficient related paper collections. After screening of 200 research papers, 70 of them met our filtering criteria, which became a part of this study. This paper presents a comprehensive review of existing parasite classification and detection methods and models in chronological order, from traditional machine learning based techniques to deep learning based techniques. In this review, we also demonstrate the summary of machine learning and deep learning methods along with dataset details, evaluation metrics, methods limitations, and future scope over the one decade. The majority of the technical publications from 2012 to the present have been examined and summarized. In addition, we have discussed the future directions and challenges of parasites classification and detection to help researchers in understanding the existing research gaps. Further, this review provides support to researchers who require an effective and comprehensive understanding of deep learning development techniques, research, and future trends in the field of parasites detection and classification.

A Conditional Generative Adversarial Network and Transfer Learning-Oriented Anomaly Classification System for Electrospun Nanofibers.

This paper proposes a generative model and transfer learning powered system for classification of Scanning Electron Microscope (SEM) images of defective nanofibers (D-NF) and nondefective nanofibers (ND-NF) produced by electrospinning (ES) process. Specifically, a conditional-Generative Adversarial Network (c-GAN) is developed to generate synthetic D-NF/ND-NF SEM images. A transfer learning-oriented strategy is also proposed. First, a Convolutional Neural Network (CNN) is pre-trained on real images. The transfer-learned CNN is trained on synthetic SEM images and validated on real ones, reporting accuracy rate up to 95.31%. The achieved encouraging results endorse the use of the proposed generative model in industrial applications as it could reduce the number of needed laboratory ES experiments that are costly and time consuming.

AIMIC: Deep Learning for Microscopic Image Classification

Background and Objective: Deep learning techniques are powerful tools for image analysis. However, the lack of programming experience makes it difficult for novice users to apply this technology. This project aims to lower the barrier for clinical users to implement deep learning methods in microscopic image classification.Methods: In this study, an out-of-the-box software, AIMIC (artificial intelligence-based microscopy image classifier), was developed for users to apply deep learning technology in a code-free manner. The platform was equipped with state-of-the-art deep learning techniques and data preprocessing approaches. Furthermore, we evaluated the built-in networks on four benchmark microscopy image datasets to assist entry-level practitioners in selecting a suitable algorithm.Results: The entire deep learning pipeline, from training a new network to inferring unseen samples using the trained model, could be implemented on the proposed platform without the need for programming. In the evaluation experiments, the ResNeXt-50-32×4d outperformed other competitor algorithms in terms of average accuracy (96.83%) and average F1-score (96.82%). In addition, the MobileNet-V2 achieved a good balance between the performance (accuracy of 95.72%) and computational cost (inference time of 0.109s for identifying one sample).Conclusions: The proposed AI platform allows people without programming experience to use artificial intelligence methods in microscopy image analysis. Besides, the ResNeXt-50-32×4d is a preferable solution for microscopic image classification, and MobileNet-V2 is most likely to be an alternative selection for the scenario when computing resources are limited.

Deep learning of rock microscopic images for intelligent lithology identification: Neural network comparison and selection

An intelligent lithology identification method is proposed based on deep learning of the rock microscopic images. Based on the characteristics of rock images in the dataset, we used Xception, MobileNet_v2, Inception_ResNet_v2, Inception_v3, Densenet121, ResNet101_v2, and ResNet-101 to develop microscopic image classification models, and then the network structures of seven different convolutional neural networks (CNNs) were compared. It shows that the multi-layer representation of rock features can be represented through convolution structures, thus better feature robustness can be achieved. For the loss function, cross-entropy is used to back propagate the weight parameters layer by layer, and the accuracy of the network is improved by frequent iterative training. We expanded a self-built dataset by using transfer learning and data augmentation. Next, accuracy (acc) and frames per second (fps) were used as the evaluation indexes to assess the accuracy and speed of model identification. The results show that the Xception-based model has the optimum performance, with an accuracy of 97.66% in the training dataset and 98.65% in the testing dataset. Furthermore, the fps of the model is 50.76, and the model is feasible to deploy under different hardware conditions and meets the requirements of rapid lithology identification. This proposed method is proved to be robust and versatile in generalization performance, and it is suitable for both geologists and engineers to identify lithology quickly.

Multi Classification of Bacterial Microscopic Images Using Inception V3

Microorganisms such as bacteria are the main cause of various infectious diseases such as cholera, botulism, gonorrhea, Lyme disease, sore throat, tuberculosis and so on. Therefore, identification and classification of bacteria is very important in the world of medicine to help experts diagnose diseases suffered by patients. However, manual identification and classification of bacteria takes a long time and a professional individual. With the help of artificial intelligence, we can effectively and efficiently classify bacteria and save a lot of time and human labor. In this study, a system was created to classify bacteria from microscopic image samples. This system uses deep learning with the transfer learning method. Inception V3 architecture was modified and retained using 108 image samples labeled with five types of bacteria, namely Acinetobacter baumanii, Escherichia coli, Neisseria gonorrhoeae, Propionibacterium acnes and Veionella. The data is then divided into training and validation using the k-fold cross validation method. Furthermore, the features that have been extracted by the model are trained with the configuration of minibatchsize 5, maxepoch 5, initiallearnrate 0.0001, and validation frequency 3. The model is then tested with data validation by conducting ten experiments and getting an average accuracy value of 94.42%.