Analysis Of Imaging Data Research Articles

Compositional brain scores capture Alzheimer's disease-specific structural brain patterns along the disease continuum.

Traditional multivariate methods for neuroimaging studies overlook the interdependent relationship between brain features. This study addresses this gap by analyzing relative brain volumetric patterns to capture how Alzheimer's disease (AD) and genetics influence brain structure along the disease continuum. This study analyzed data from participants across the AD continuum from the Alzheimer's and Families (ALFA) and Alzheimer's Disease Neuroimaging Initiative (ADNI) studies. Compositional data analysis (CoDA) was exploited to examine relative brain volumetric variations that (1) were linked to different AD stages compared to cognitively unimpaired amyloid-β-negative (CU A-) individuals and (2) varied by AD genetic risk. Disease stage-specific compositional brain scores were identified, differentiating CU A- individuals from those in more advanced stages. Genetic risk-stratified models revealed a broader genetic landscape affecting brain morphology in AD, beyond the well-known apolipoprotein E ε4 allele. CoDA emerges as an alternative multivariate framework to deepen understanding of AD-related structural changes and support targeted interventions for those at higher genetic risk. Compositional data analysis (CoDA) revealed the relative variation of brain region volumes, captured in compositional brain scores, capable of discerning between cognitively unimpaired amyloid-β-negative individuals and subjects within other disease-stage groups along the Alzheimer's disease (AD) continuum. CoDA also uncovered the genetic vulnerability of specific brain regions at each stage of the disease along the continuum. CoDA is capable of integrating magnetic resonance imaging data from two different cohorts without stringent requirements for harmonization. This translates as an advantage, compared to traditional methods, and strengthens the reliability of cross-study comparisons by standardizing the data despite different labeling agreements, facilitating collaborative and large-scale research. The algorithm is sensitive to AD-specific effects, as the main compositional brain scores display little overlap with the age-specific compositional brain score. CoDA provides a more accurate analysis of brain imaging data addressing its compositional nature, which can influence the development of targeted approaches, opening new avenues for enhancing brain health.

MsiFlow: automated workflows for reproducible and scalable multimodal mass spectrometry imaging and microscopy data analysis

Multimodal imaging by matrix-assisted laser desorption ionisation mass spectrometry imaging (MALDI MSI) and microscopy holds potential for understanding pathological mechanisms by mapping molecular signatures from the tissue microenvironment to specific cell populations. However, existing software solutions for MALDI MSI data analysis are incomplete, require programming skills and contain laborious manual steps, hindering broadly applicable, reproducible, and high-throughput analysis to generate impactful biological discoveries. Here, we present msiFlow, an accessible open-source, platform-independent and vendor-neutral software for end-to-end, high-throughput, transparent and reproducible analysis of multimodal imaging data. msiFlow integrates all necessary steps from raw data import to analytical visualisation along with state-of-the-art and self-developed algorithms into automated workflows. Using msiFlow, we unravel the molecular heterogeneity of leukocytes in infected tissues by spatial regulation of ether-linked phospholipids containing arachidonic acid. We anticipate that msiFlow will facilitate the broad applicability of MSI in multimodal imaging to uncover context-dependent cellular regulations in disease states.

Early Diagnosis of Brain Tumor Using AI and Mathematical Modeling

Timely detection of Brain Tumors is crucial for improving patient outcomes and reducing mortality rates. This study explores the significance of early diagnosis and investigates the role of various diagnostic methods. Leveraging advancements in medical imaging, including MRI, CT scans, and PET scans, alongside blood tests measuring biomarkers such as glial fibrillary acidic protein (GFAP) and neuron-specific enolase (NSE), facilitates early detection of Brain Tumors. Furthermore, emerging technologies like artificial intelligence algorithms enhance the accuracy and efficiency of diagnosis by analyzing imaging data and identifying subtle abnormalities. Early detection not only offers patients more treatment options but also improves their prognosis. Therefore, it is crucial to raise awareness about the importance of regular screenings and symptom awareness to ensure early detection and timely intervention for Brain Tumors. Keywords: Brain tumor, Early detection, Artificial intelligence, Mathematical modeling, Medical imaging, Mammography, liver MRI, Machine learning, Deep learning, Screening strategies.

Automated cell annotation in multi-cell images using an improved CRF_ID algorithm.

Cell identification is an important yet difficult process in data analysis of biological images. Previously, we developed an automated cell identification method called CRF_ID and demonstrated its high performance in Caenorhabditis elegans whole-brain images (Chaudhary et al., 2021). However, because the method was optimized for whole-brain imaging, comparable performance could not be guaranteed for application in commonly used C. elegans multi-cell images that display a subpopulation of cells. Here, we present an advancement, CRF_ID 2.0, that expands the generalizability of the method to multi-cell imaging beyond whole-brain imaging. To illustrate the application of the advance, we show the characterization of CRF_ID 2.0 in multi-cell imaging and cell-specific gene expression analysis in C. elegans. This work demonstrates that high-accuracy automated cell annotation in multi-cell imaging can expedite cell identification and reduce its subjectivity in C. elegans and potentially other biological images of various origins.

Generative Artificial Intelligence in Anatomic Pathology.

Generative artificial intelligence (AI) has emerged as a transformative force in various fields, including anatomic pathology, where it offers the potential to significantly enhance diagnostic accuracy, workflow efficiency, and research capabilities. To explore the applications, benefits, and challenges of generative AI in anatomic pathology, with a focus on its impact on diagnostic processes, workflow efficiency, education, and research. A comprehensive review of current literature and recent advancements in the application of generative AI within anatomic pathology, categorized into unimodal and multimodal applications, and evaluated for clinical utility, ethical considerations, and future potential. Generative AI demonstrates significant promise in various domains of anatomic pathology, including diagnostic accuracy enhanced through AI-driven image analysis, virtual staining, and synthetic data generation; workflow efficiency, with potential for improvement by automating routine tasks, quality control, and reflex testing; education and research, facilitated by AI-generated educational content, synthetic histology images, and advanced data analysis methods; and clinical integration, with preliminary surveys indicating cautious optimism for nondiagnostic AI tasks and growing engagement in academic settings. Ethical and practical challenges require being addressed by rigorous validation, prompt engineering, federated learning, and synthetic data generation to help ensure trustworthy, reliable, and unbiased AI applications. Generative AI can potentially revolutionize anatomic pathology, enhancing diagnostic accuracy, improving workflow efficiency, and advancing education and research. Successful integration into clinical practice will require continued interdisciplinary collaboration, careful validation, and adherence to ethical standards to ensure the benefits of AI are realized while maintaining the highest standards of patient care.

Light sheet fluorescence microscopy for monitoring drug delivery: Unlocking the developmental phases of embryos.

Light sheet fluorescence microscopy (LSFM) has emerged as a transformative imaging technique in the study of drug delivery and embryonic development, offering high-resolution, real-time visualization with minimal phototoxicity. This review examines the application of LSFM in tracking drug pharmacokinetics, tissue-specific targeting, and drug efficacy during critical phases of embryonic development. Recent advancements in fluorescent labeling and machine learning integration have enabled more precise monitoring of drug release, distribution, and interaction with developing tissues. The ability of LSFM to capture long-term dynamics at single-cell resolution has revolutionized drug discovery, especially in nanomedicine and targeted therapies. By integrating LSFM with multimodal imaging and AI-driven data analysis, researchers are now better equipped to explore complex biological processes and optimize drug delivery in a highly controlled, minimally invasive manner. Finally, the review highlights the pivotal role of LSFM in advancing drug delivery research, addressing existing challenges, and unlocking new frontiers in clinical applications.

Innovative segmentation technique for aerial power lines via amplitude stretching transform

Accurate segmentation of power line targets helps quickly locate faults, evaluate line conditions, and provides key image data support and analysis for the safe and stable operation of the power system.The aerial power line in segmentation due to the target is small, and the imaging reflected energy is weak, so the Unmanned Aerial Vehicle (UAV) aerial power line image is very susceptible to the interference of the environment line elements and noise, resulting in the detection of the power line target in the image of the defective, intermittent, straight line interferences and other low accuracy and real-time efficiency is not high. For this reason, this paper designs a pure amplitude stretching kernel function to form a Fourier amplitude vector field and uses this amplitude vector field to implement the stretching transformation of the amplitude field of the aerial power line image, so that the angular field after the Fourier inverse transformation can better react to the spatial domain line targets, and finally, after the Relative Total Variation (RTV) processing, the power line can be well detected. The proposed algorithm is compared with the main power line segmentation algorithms, such as Region Convolutional Neural Networks(R-CNN) and Phase Stretch Transform(PST). The average values of evaluation indicators PPA, MMPA and MMIoU of the image segmentation results of the proposed algorithm reach 0.96, 0.96 and 0.95 respectively, and the average time lag of detection is less than 0.2s, indicating that the accuracy and real-time performance of the segmentation results of the proposed algorithm are significantly better than those of the above algorithms.

CalciumZero: a toolbox for fluorescence calcium imaging on iPSC derived brain organoids

Calcium plays an important role in regulating various neuronal activities in human brains. Investigating the dynamics of the calcium level in neurons is essential not just for understanding the pathophysiology of neuropsychiatric disorders but also as a quantitative gauge to evaluate the influence of drugs on neuron activities. Accessing human brain tissue to study neuron activities has historically been challenging due to ethical concerns. However, a significant breakthrough in the field has emerged with the advent of utilizing patient-derived human induced pluripotent stem cells (iPSCs) to culture neurons and develop brain organoids. This innovative approach provides a promising modeling system to overcome these critical obstacles. Many robust calcium imaging analysis tools have been developed for calcium activity analysis. However, most of the tools are designed for calcium signal detection only. There are limited choices for in-depth downstream applications, particularly in discerning differences between patient and normal calcium dynamics and their responses to drug treatment obtained from human iPSC-based models. Moreover, end-user researchers usually face a considerable challenge in mastering the entire analysis procedure and obtaining critical outputs due to the steep learning curve associated with these available tools. Therefore, we developed CalciumZero, a user-friendly toolbox to satisfy the unmet needs in calcium activity studies in human iPSC-based 3D-organoid/neurosphere models. CalciumZero includes a graphical user interface (GUI), which provides end-user iconic visualization and smooth adjustments on parameter tuning. It streamlines the entire analysis process, offering full automation with just one click after parameter optimization. In addition, it includes supplementary features to statistically evaluate the impact on disease etiology and the detection of drug candidate effects on calcium activities. These evaluations will enhance the analysis of imaging data obtained from patient iPSC-derived brain organoid/neurosphere models, providing a more comprehensive understanding of the results.

ADC histogram analysis of tumor-infiltrating CD8+ T cell levels in meningioma.

To investigate the value of preoperative MRI features and ADC histogram analysis for evaluating tumor-infiltrating CD8+ T cells in meningiomas. In this single-center cross-sectional study, we conducted a retrospective analysis of clinical, imaging, and pathological data from 84 patients with meningioma and performed immunohistochemical staining to quantitatively evaluate CD8+ T cells. Using X-Tile software, we divided the patients into high-and low-CD8+ T cells groups based on cut-off values. Furthermore, we compared the clinical and MRI features between the two groups and assessed the predictive value of significant parameters by plotting ROC curves. Additionally, Spearman's analysis was used to examine the association between ADC histogram parameters and CD8+ T cells. The level of tumor-infiltrating CD8+ T cells was found to have a negative correlation with recurrence in patients with meningiomas (r=-0.235, p = 0.031). No statistically significant differences were found in clinical and conventional MRI features between the two groups (all p > 0.05). Conversely, among the ADC histogram parameters, the coefficient of variation (CV), Perc.01, Perc.05, Perc.10, and Perc.25 showed statistically significant differences between the two groups (all p < 0.05) and combined ADC histogram parameters had the highest AUC (0.791; 95%CI (0.689-0.872)). Additionally, we observed a positive correlation between Perc.01, Perc.05, Perc.10 and CD8+ T cells (p < 0.05), the CV and variance was negatively correlated with the levels of CD8+ T cells (p < 0.05). ADC histogram analysis can be used as an imaging tool to preoperatively assess CD8+ T cells in patients with meningioma, and found a certain correlation between them.

Time-Lapse Imaging to Analyze Cell Fate in Response to Antimitotics.

Time-lapse imaging is a powerful technique widely used in Cellular and Molecular Biology to capture and analyze the dynamic processes of living cells over specific time periods. Particularly, in the study of cell division, under normal conditions or after drug treatments, this methodology can provide essential data that is impossible to obtain from conventional cell-fixed-based assays. In the context of evaluating cell fate after antimitotic drug treatment, time-lapse imaging provides valuable insights into the mechanisms of action of these compounds. Antimitotic drugs belong to a class of anticancer compounds that interfere with mitosis by targeting microtubules or specific kinases and molecular motors involved in the mitotic apparatus. This chapter aims to highlight the significant advantage of using time-lapse microscopy to assess the effects of antimitotic drugs on cell behavior and fates. We describe the full protocol for time-lapse imaging, using the human lung adenocarcinoma A549 cell line as a model, after exposure to the antimitotic BI2536, a potent inhibitor of polo-like kinase 1 (PLK1). This description includes software for imaging acquisition and data analysis.

Compare Noise Robust Least-Squares Method with Other Methods for Estimation of the Parameters of Frechet Distribution and Neutrosophic Generalization

The Frechet distribution is a versatile probability distribution that is used within a loose range in many important statistical fields, such as image processing, data analysis, and pattern recognition. It aims to explore and study the estimation of the parameters of the Frechet distribution using the noise-robust least squares method, as in the research paper, and it also has uses. There are many real-world scenarios. It is known that there is a growing challenge in estimating the parameter because of the noisy data. Depending on rigorous simulations and experimental analysis, we provide a novel powerful way to estimate the parameters for the Frechet Distribution Robust Least Squares approach to be flexible. Also, the results approach of this work will be very helpful in estimating the Frechet distribution parameters for diverse statistical applications. Also, we generalize our results to include the generalized neutrosophic case of this distribution dealing with neutrosophic numbers.

Differences in spatiotemporal dynamics for processing specific semantic categories: An EEG study

Semantic processing is an essential mechanism in human language comprehension and has profound implications for speech brain-computer interface technologies. Despite recent advances in brain imaging techniques and data analysis algorithms, the mechanisms underlying human brain semantic representations remain a topic of debate, specifically whether this occurs through the activation of selectively separated cortical regions or via a network of distributed and overlapping regions. This study investigates spatiotemporal neural representation during the perception of semantic words related to faces, numbers, and animals using electroencephalography. Source‐level analysis focuses on contrasting neural responses to different semantic categories. Critical intervals used in the source contrast analysis are defined using the peak duration of global field power. Effective connectivity, determined through a causality analysis of brain regions activated for semantic processing, is explored. The findings reveal the necessity of a distributed network of regions for processing specific semantic categories and provide evidence suggesting the existence of a neural substrate for semantic representations.

Accurate and fast segmentation of filaments and membranes in micrographs and tomograms with TARDIS.

It is now possible to generate large volumes of high-quality images of biomolecules at near-atomic resolution and in near-native states using cryogenic electron microscopy/electron tomography (Cryo-EM/ET). However, the precise annotation of structures like filaments and membranes remains a major barrier towards applying these methods in high-throughput. To address this, we present TARDIS (Transformer-based Rapid Dimensionless Instance Segmentation), a machine-learning framework for fast and accurate annotation of micrographs and tomograms. TARDIS combines deep learning for semantic segmentation with a novel geometric model for precise instance segmentation of various macromolecules. We develop pre-trained models within TARDIS for segmenting microtubules and membranes, demonstrating high accuracy across multiple modalities and resolutions, enabling segmentation of over 13,000 tomograms from the CZI Cryo-Electron Tomography data portal. As a modular framework, TARDIS can be extended to new structures and imaging modalities with minimal modification. TARDIS is open-source and freely available at https://github.com/SMLC-NYSBC/TARDIS, and accelerates analysis of high-resolution biomolecular structural imaging data.

Comparative Analysis of Conventional CNN v’s ImageNet Pretrained ResNet in Medical Image Classification

Convolutional Neural Networks (CNNs) are the prevalent technology in computer vision and have become increasingly popular for medical imaging data classification and analysis. In this field, due to the scarcity of medical data, pretrained ResNets on ImageNet can be considered a suitable first approach. This paper examines the medical imaging classification accuracy of conventional basic custom CNNs compared to ImageNet pretrained ResNets on various medical datasets in an effort to give more information about the importance of medical data and its preprocessing techniques for disease studies. Microscope-extracted cytological images were examined along with chest X-rays, MRI brain scans, and melanoma photographs. The medical images were examined in various sets, class combinations, and resolutions. Augmented image datasets and asymmetrical training and validation splits among the classes were also examined. Models were developed after they were tested and fine-tuned with respect to their network size, parameter values and network methods, image resolution, size of dataset, multitude, and genre of class. Overfitting was also examined, and comparative studies regarding the computational cost of different models were performed. The models achieved high accuracy in image classification that varies depending on the dataset and can be easily incorporated in future over-the-internet medical decision-supporting (telemedicine) environments. In addition, it appeared that conventional basic custom CNN overperformed ImageNet pretrained ResNets. The obtained results indicate the importance of utilizing medical image data as a testbed for improvements in CNN classification performance and the possibility of using CNNs and data preprocessing techniques for disease studies.

Recent Advances in Big Medical Image Data Analysis Through Deep Learning and Cloud Computing

This comprehensive study investigates the integration of cloud computing and deep learning technologies in medical data analysis, focusing on their combined effects on healthcare delivery and patient outcomes. Through a methodical examination of implementation instances at various healthcare facilities, we investigate how well these technologies manage a variety of medical data sources, such as wearable device data, medical imaging data, and electronic health records (EHRs). Our research demonstrates significant improvements in diagnostic accuracy (15–20% average increase) and operational efficiency (60% reduction in processing time) when utilizing cloud-based deep learning systems. We found that healthcare organizations implementing phased deployment approaches achieved 90% successful integration rates, while hybrid cloud architectures improved regulatory compliance by 50%. This study also revealed critical challenges, with 35% of implementations facing data integration issues and 5% experiencing security breaches. Through empirical analysis, we propose a structured implementation framework that addresses these challenges while maintaining high performance standards. Our findings indicate that federated learning techniques retain 95% model accuracy while enhancing privacy protection, and edge computing reduces latency by 40% in real-time processing. By offering quantitative proof of the advantages and difficulties of combining deep learning and cloud computing in medical data analysis, as well as useful recommendations for healthcare organizations seeking technological transformation, this study adds to the expanding body of knowledge on healthcare digitalization.

Diagnostic errors in assessing magnetic resonance imaging semiotics of primary extracerebral tumors: a case series

Primary extracerebral tumors are represented by benign and malignant neoplasms of the meninges and cranial nerves. Their presurgical differential diagnosis is based on the analysis of magnetic resonance imaging semiotics. The critically significant aspects for classifying tumors of this group include the following: neoplasm structure, contrast enhancement type, delimiting from the brain tissue, and relationship with the meninges or cranial nerves. Differential diagnosis of extracerebral tumors based on visual analysis of magnetic resonance imaging data is generally not challenging because most tumors have typical magnetic resonance imaging semiotics. However, in cases with atypical magnetic resonance imaging signs, reliable differentiation of tumors can be challenging. Moreover, the greatest complexity is the differentiation of meningioma grades, distinction between solitary fibrous tumors and meningiomas, and identification of the tumor type when localized in the cerebellopontine angles. The case series presented the most typical features leading to errors in the differential diagnosis of primary extracerebral tumors. All the presented tumors were verified with postsurgical histological examination. The analysis of the case reports demonstrates that a review of the combined semiotic signs can lower the number of diagnostic errors.

Constrained Independent Vector Analysis with Reference for Multi-Subject fMRI Analysis.

Independent component analysis (ICA) is now a widely used solution for the analysis of multi-subject functional magnetic resonance imaging (fMRI) data. Independent vector analysis (IVA) generalizes ICA to multiple datasets (multi-subject data). Along with higher-order statistical information in ICA, it leverages the statistical dependence across the datasets as an additional type of statistical diversity. As such, IVA preserves variability in the estimation of single-subject maps but its performance might suffer when the number of datasets increases. Constrained IVA is an effective way to bypass computational issues and improve the quality of separation by incorporating available prior information. Existing constrained IVA approaches often rely on user-defined threshold values to define the constraints. However, an improperly selected threshold can have a negative impact on the final results. This paper proposes two novel methods for constrained IVA: one using an adaptive-reverse scheme to select variable thresholds for the constraints and a second one based on a threshold-free formulation by leveraging the unique structure of IVA. Notably, the proposed algorithms do not require all components to be constrained, utilizing free components to model interferences and components that might not be in the reference set. We demonstrate that our solutions provide an attractive solution to multi-subject fMRI analysis both by simulations and through analysis of resting state fMRI data collected from 98 subjects - the highest number of subjects ever used by IVA algorithms. Our results show that both proposed approaches obtain significantly better separation quality and model match while providing computationally efficient and highly reproducible solutions.

The application of “transfer learning” in optical microscopy: The petrographic classification of opaque minerals

Abstract The analysis of optical microscopic image data is crucial for the identification and characterization of mineral phases and, thus, directly relevant to the subsequent methodology selections of further detailed petrological exploration. Here, we present a novel application of Swin Transformer, a deep learning algorithm to classify mineral phases such as arsenopyrite, chalcopyrite, gold, pyrite, and stibnite in images captured by optical microscopy. To speed up the training process and improve the generalization capabilities of the investigated model, we adopt the “transfer learning” paradigm by pre-training the algorithm using a large, general-purpose image data set named ImageNet-1k. Furthermore, we compare the performances of the Swin Transformer with those of two well-established Convolutional Neural Networks (CNNs) named MobileNetv2 and ResNet50, respectively. Our results highlight a maximum accuracy of 0.92 for the Swin Transformer, outperforming the CNNs. To provide an interpretation of the trained models, we apply the so-called Class Activation Map (CAM), which indicates a strong global feature extraction ability of the Swin Transformer metal mineral classifier that focuses on distinctive (e.g., colors) and microstructural (e.g., edge shapes) features. The results demonstrate that the deep learning approach can accurately extract all available attributes, which reveals the potential to assist in data exploration and provides an opportunity to carry out spatial quantization at a large scale (centimeters-millimeters). Simultaneously, boosting the learning processes with pre-trained weights can accurately capture relevant attributes in mineral classification, revealing the potential for application in mineralogy and petrology, as well as enabling its use in resource explorations.

Comparative Analysis of RGB and Multispectral UAV Image Data for Leaf Area Index Estimation of Sweet Potato

This study aims to evaluate whether the utilization of a broader array of image types and the extraction of plant canopy using RGB and multispectral cameras mounted on unmanned aerial vehicles (UAVs) can improve leaf area index (LAI) estimation accuracy. The accuracy of LAI estimation for sweet potato was compared across different model types—mono-regression models based on a single image, multivariate regression models based on a single type of image, and multivariate regression models based on multiple image types—by employing 10 types of images based on color, morphological indices, vegetation indices (VIs) from RGB cameras, reflectance, VIs from multispectral cameras, and each of their canopy part image. For each regression model, we compared the estimation accuracy based on the whole image with that based on the canopy part image. For the mono-regression model, EVI2 from the whole image exhibited the lowest test root mean squared error (RMSE) of 0.403, which is attributed to EVI2’s capacity to mitigate the effects of spectral saturation. Contrary to the models with multiple image types that demonstrated improved accuracy, the multivariate regression models based on a single image type did not enhance estimation accuracy; this shows that the use of multiple image types improves the LAI estimation accuracy owing to the synergy of different spectral information from various image types. Plant canopy extraction did not enhance the estimation accuracy in mono-regression models or multivariate regression models based on a single image type; however, it improved the accuracy of multivariate regression models based on multiple image types, provided the image types were appropriately combined. The highest estimation accuracy was achieved by a partial least squares regression model based on VI (from the RGB camera) and reflectance (from the multispectral camera) of the canopy part (test R2 = 0.887, test RMSE = 0.351), suggesting that this is an effective approach for LAI estimation in sweet potato (cultivar Beniharuka). Overall, this study demonstrates that an optimal combination of image type and plant canopy extraction can enhance LAI estimation accuracy in UAV-based monitoring.

Revolutionizing Epithelial Differentiability Analysis in Small Airway-on-a-Chip Models Using Label-Free Imaging and Computational Techniques.

Organ-on-a-chip (OOC) devices mimic human organs, which can be used for many different applications, including drug development, environmental toxicology, disease models, and physiological assessment. Image data acquisition and analysis from these chips are crucial for advancing research in the field. In this study, we propose a label-free morphology imaging platform compatible with the small airway-on-a-chip system. By integrating deep learning and image recognition techniques, we aim to analyze the differentiability of human small airway epithelial cells (HSAECs). Utilizing cell imaging on day 3 of culture, our approach accurately predicts the differentiability of HSAECs after 4 weeks of incubation. This breakthrough significantly enhances the efficiency and stability of establishing small airway-on-a-chip models. To further enhance our analysis capabilities, we have developed a customized MATLAB program capable of automatically processing ciliated cell beating images and calculating the beating frequency. This program enables continuous monitoring of ciliary beating activity. Additionally, we have introduced an automated fluorescent particle tracking system to evaluate the integrity of mucociliary clearance and validate the accuracy of our deep learning predictions. The integration of deep learning, label-free imaging, and advanced image analysis techniques represents a significant advancement in the fields of drug testing and physiological assessment. This innovative approach offers unprecedented insights into the functioning of the small airway epithelium, empowering researchers with a powerful tool to study respiratory physiology and develop targeted interventions.