Raw Images Research Articles

Spot-Based Global Registration for Subpixel Stitching of Single-Molecule Resolution Images for Tissue-Scale Spatial Transcriptomics.

Single-molecule imaging at the tissue scale has revolutionized our understanding of biology by providing unprecedented insight into the molecular expression of individual cells and their spatial organization within tissues. However, achieving precise image stitching at the single-molecule level remains a challenge, primarily due to heterogeneous background signals and dim labeling signals in single-molecule images. This paper introduces Spot-Based Global Registration (SBGR), a novel strategy that shifts the focus from raw images to identified molecular spots for high-resolution image alignment. The use of spot-based data enables straightforward and robust evaluation of the credibility of estimated translations and stitching performance. The method outperforms existing image-based stitching methods, achieving subpixel accuracy (83 ± 36 nm) with exceptional consistency. Furthermore, SBGR incorporates a mechanism to surgically remove duplicate spots in overlapping regions, maximizing information recovery from duplicate measurements. In conclusion, SBGR emerges as a robust and accurate solution for stitching single-molecule resolution images in tissue-scale spatial transcriptomics, offering versatility and potential for high-resolution spatial analysis.

Using explainable AI to investigate electrocardiogram changes during healthy aging-From expert features to raw signals.

Cardiovascular diseases remain the leading global cause of mortality. Age is an important covariate whose effect is most easily investigated in a healthy cohort to properly distinguish the former from disease-related changes. Traditionally, most of such insights have been drawn from the analysis of electrocardiogram (ECG) feature changes in individuals as they age. However, these features, while informative, may potentially obscure underlying data relationships. In this paper we present the following contributions: (1) We employ a deep-learning model and a tree-based model to analyze ECG data from a robust dataset of healthy individuals across varying ages in both raw signals and ECG feature format. (2) We use explainable AI methods to identify the most discriminative ECG features across age groups.(3) Our analysis with tree-based classifiers reveals age-related declines in inferred breathing rates and identifies notably high SDANN values as indicative of elderly individuals, distinguishing them from younger adults. (4) Furthermore, the deep-learning model underscores the pivotal role of the P-wave in age predictions across all age groups, suggesting potential changes in the distribution of different P-wave types with age. These findings shed new light on age-related ECG changes, offering insights that transcend traditional feature-based approaches.

Automated detection and classification of concealed objects using infrared thermography and convolutional neural networks

This paper presents a study on the effectiveness of a convolutional neural network (CNN) in classifying infrared images for security scanning. Infrared thermography was explored as a non-invasive security scanner for stand-off and walk-through concealed object detection. Heat generated by human subjects radiates off the clothing surface, allowing detection by an infrared camera. However, infrared lacks in penetration capability compared to longer electromagnetic waves, leading to less obvious visuals on the clothing surface. ResNet-50 was used as the CNN model to automate the classification process of thermal images. The ImageNet database was used to pre-train the model, which was further fine-tuned using infrared images obtained from experiments. Four image pre-processing approaches were explored, i.e., raw infrared image, subject cropped region-of-interest (ROI) image, K-means, and Fuzzy-c clustered images. All these approaches were evaluated using the receiver operating characteristic curve on an internal holdout set, with an area-under-the-curve of 0.8923, 0.9256, 0.9485, and 0.9669 for the raw image, ROI cropped, K-means, and Fuzzy-c models, respectively. The CNN models trained using various image pre-processing approaches suggest that the prediction performance can be improved by the removal of non-decision relevant information and the visual highlighting of features.

FastDFE: Fast and Flexible Inference of the Distribution ofFitness Effects.

Estimating the distribution of fitness effects (DFE) of new mutations is of fundamental importance in evolutionary biology, ecology, and conservation. However, existing methods for DFE estimation suffer from limitations, such as slow computation speed and limited scalability. To address these issues, we introduce fastDFE, a Python-based software package, offering fast, and flexible DFE inference from site-frequency spectrum (SFS) data. Apart from providing efficient joint inference of multiple DFEs that share parameters, it offers the feature of introducing genomic covariates that influence the DFEs and testing their significance. To further simplify usage, fastDFE is equipped with comprehensive VCF-to-SFS parsing utilities. These include options for site filtering and stratification, as well as site-degeneracy annotation and probabilistic ancestral-allele inference. fastDFE thereby covers the entire workflow of DFE inference from the moment of acquiring a raw VCF file. Despite its Python foundation, fastDFE incorporates a full R interface, including native R visualization capabilities. The package is comprehensively tested and documented at fastdfe.readthedocs.io.

Application of Convolutional Neural Networks for Early Detection and Classification of Alzheimer's disease from MRI Images

This study investigates the application of convolutional neural networks (CNNs) and traditional machine learning algorithms for the early detection and classification of Alzheimer's disease (AD) using brain Magnetic Resonance Imaging (MRI) data. We compare the performance of CNNs with Support Vector Machines (SVM), Random Forest (RF), and Gradient Boosting Machines (GBM) on a dataset comprising MRI images from AD patients and healthy controls. Results show that CNNs achieved the highest accuracy (90.2%) and area under the receiver operating characteristic curve (AUC-ROC) of 0.95, outperforming SVM, RF, and GBM. The CNN model also exhibited high sensitivity (87.5%) and specificity (92.6%) in distinguishing between AD patients and healthy controls. These findings highlight the effectiveness of CNN-based approaches in leveraging raw MRI images for accurate and early detection of AD.

Two-dimensional time-resolved scintillating sheet monitoring of proton pencil beam scanning FLASH mouse irradiations.

Dosimetry in pre-clinical FLASH studies is essential for understanding the beam delivery conditions that trigger the FLASH effect. Resolving the spatial and temporal characteristics of proton pencil beam scanning (PBS) irradiations with ultra-high dose rates (UHDR) requires a detector with high spatial and temporal resolution. To implement a novel camera-based system for time-resolved two-dimensional (2D) monitoring and apply it in vivo during pre-clinical proton PBS mouse irradiations. Time-resolved 2D beam monitoring was performed with a scintillation imaging system consisting of a 1mm thick transparent scintillating sheet, imaged by a CMOS camera. The sheet was placed in a water bath perpendicular to a horizontal PBS proton beam axis. The scintillation light was reflected through a system of mirrors and captured by the camera with 500 frames per second (fps) for UHDR and 4 fps for conventional dose rates. The raw images were background subtracted, geometrically transformed, flat field corrected, and spatially filtered. The system was used for 2D spot and field profile measurements and compared to radiochromic films. Furthermore, spot positions were measured for UHDR irradiations. The measured spot positions were compared to the planned positions and the relative instantaneous dose rate to equivalent fiber-coupled point scintillator measurements. For in vivo application, the scintillating sheet was placed 1cm upstream the right hind leg of non-anaesthetized mice submerged in the water bath. The mouse leg and sheet were both placed in a 5cm wide spread-out Bragg peak formed from the mono-energetic proton beam by a 2D range modulator. The mouse leg position within the field was identified for both conventional and FLASH irradiations. For the conventional irradiations, the mouse foot position was tracked throughout the beam delivery, which took place through repainting. For FLASH irradiations, the delivered spot positions and relative instantaneous dose rate were measured. The pixel size was 0.1mm for all measurements. The spot and field profiles measured with the scintillating sheet agreed with radiochromic films within 0.4mm. The standard deviation between measured and planned spot positions was 0.26mm and 0.35mm in the horizontal and vertical direction, respectively. The measured relative instantaneous dose rate showed a linear relation with the fiber-coupled scintillator measurements. For in vivo use, the leg position within the field varied between mice, and leg movement up to 3mm was detected during the prolonged conventional irradiations. The scintillation imaging system allowed for monitoring of UHDR proton PBS delivery in vivo with 0.1mm pixel size and 2ms temporal resolution. The feasibility of instantaneous dose rate measurements was demonstrated, and the system was used for validation of the mouse leg position within the field.

High-resolution cell imaging using white light phase shifting interferometry and iterative phase deconvolution.

An optimization algorithm is presented for the deconvolution of a complex field to improve the resolution and accuracy of quantitative phase imaging (QPI). A high-resolution phase map can be recovered by solving a constrained optimization problem of deconvolution using a complex gradient operator. The method is demonstrated on phase measurements of samples using a white light based phase shifting interferometry (WLPSI) method. The application of the algorithm on real and simulated objects shows a significant resolution and contrast improvement. Experiments performed on Escherichia coli bacterium have revealed its sub-cellular structures that were not visible in the raw WLPSI images obtained using a five phase shifting method. These features can give valuable insights into the structures and functioning of biological cells. The algorithm is simple in implementation and can be incorporated into other QPI modalities .

Quantitative assessment of cement bridges and voids in cement-stabilized permeable base materials using a mask R-CNN-based CT image segmentation strategy

Accurate extraction of aggregates, cement bridges, and voids in permeable cement-stabilized base materials (PCBM) is a challenging task. This paper proposed a segmentation and reconstruction paradigm for raw CT slice images using a mask regional convolutional neural network (Mask R-CNN). The results demonstrate that the proposed method effectively segments and reconstructs unclear raw CT slice images, providing a convenient and advanced tool for accurately assessing the meso-scale structures of cement bridges and voids in PCBM prepared with different combinations of cement content and compaction force. An increased cement content or compaction force results in the migration of cement paste, leading to thicker localized cement bridges and blockage of permeable pore space. This phenomenon further disrupts the homogeneous distribution of cement bridges and voids. An increased compaction force tends to elongate and flatten pore throats, while an increased cement content may clog voids, rendering them impermeable. The vertical permeability of the PCBM is significantly greater than the horizontal permeability, which should be considered during material design optimization. The proposed method could serve as a reference for microscopic analysis and structural design optimization of PCBM.

A Deformable Constraint Transport Network for Optimal Aortic Segmentation From CT Images.

Aortic segmentation from computed tomography (CT) is crucial for facilitating aortic intervention, as it enables clinicians to visualize aortic anatomy for diagnosis and measurement. However, aortic segmentation faces the challenge of variable geometry in space, as the geometric diversity of different diseases and the geometric transformations that occur between raw and measured images. Existing constraint-based methods can potentially solve the challenge, but they are hindered by two key issues: inaccurate definition of properties and inappropriate topology of transformation in space. In this paper, we propose a deformable constraint transport network (DCTN). The DCTN adaptively extracts aortic features to define intra-image constrained properties and guides topological implementation in space to constrain inter-image geometric transformation between raw and curved planar reformation (CPR) images. The DCTN contains a deformable attention extractor, a geometry-aware decoder and an optimal transport guider. The extractor generates variable patches that preserve semantic integrity and long-range dependency in long-sequence images. The decoder enhances the perception of geometric texture and semantic features, particularly for low-intensity aortic coarctation and false lumen, which removes background interference. The guider explores the geometric discrepancies between raw and CPR images, constructs probability distributions of discrepancies, and matches them with inter-image transformation to guide geometric topology in space. Experimental studies on 267 aortic subjects and four public datasets show the superiority of our DCTN over 23 methods. The results demonstrate DCTN's advantages in aortic segmentation for different types of aortic disease, for different aortic segments, and in the measurement of clinical indexes.

Non-contact, portable, and stand-off infrared thermal imager for security scanning applications

In this article, we demonstrated the physical application of a portable infrared (IR) security scanning system for the non-contact and stand-off detection of target objects concealed underneath clothing. Such a system combines IR imaging and transfer learning with convolutional neural networks (CNNs) to enhance the detection of weak thermal signals and automate the classification of IR images. A mid-wavelength IR detector was used to record the real-time heat emitted from the clothing surface by human subjects. Concealed objects reduce the transmissivity of IR radiation from the body to the clothing surface, generally showing lower IR intensity compared to regions without objects. Due to limited resources for training data, the transfer learning approach was applied by fine-tuning a pre-trained CNN ResNet-50 model using the ImageNet database. Two image types were investigated here, i.e., raw thermal and Fuzzy-c clustered images. Receiver operating characteristic curves were built using a holdout set, showing an area-under-the-curve of 0.8934 and 0.9681 for the raw and Fuzzy-c clustered image models, respectively. The gradient-weighted class activation mapping visualization method was used to improve target identification, showing an accurate prediction of the object area. It was also found that complex clothing, such as those composed of materials of different transmissivity, could mislead the model in classification. The proposed IR-based detector has shown potential as a non-contact, stand-off security scanning system that can be deployed in diverse locations and ensure the safety of civilians.

Raw Image Based Over-Exposure Correction Using Channel-Guidance Strategy

Raw Image Based Over-Exposure Correction Using Channel-Guidance Strategy

ANOMALY XPERT: A Deep Learning Approach to Anomaly Detection

Abstract: In response to increasing crime rates, organizations are deploying surveillance systems with CCTV cameras to detect suspicious activities autonomously. This paper proposes an automated system using transfer learning-based CNN models to track and classify activities like 'Shoplifting,' 'Robbery,' or 'Break-In' in real-time CCTV footage. The framework processes raw camera data, detects objects, tracks activities, and classifies them, generating alerts for authorized personnel. Leveraging transfer learning enhances the precision and effectiveness of the CNN model in identifying security threats. Preliminary evaluations demonstrate promising outcomes, yet additional investigation is warranted to address obstacles such as occlusions and lighting variations. Overall, this system offers a proactive security solution for retail environments, ensuring timely detection and intervention against potential security breaches

A lab-centric, workflow-based data management system for environmental DNA research

The adoption of environmental DNA approaches as a standard tool for biodiversity monitoring leads to the increase in the number of eDNA-based species occurrence records; however, considerable disparity remains in the nature and quality of associated information, much of it unpublished and/or poorly parametrised. A robust system for tracking biological materials from their point of origin through laboratory analyses is required to connect inferred taxon occurrences with analytical history and provenance data. The bulk of eDNA research is currently driven by small-scale operations where the tasks of digitisation, organisation and cross-referencing field records with laboratory analytical data and biomaterial sample location, are often performed manually and disconnected. We present an integrative, full-stack data management solution that provides a structured ontological concept, a minimalist data schema for eDNA research and a software application prototype designed to facilitate real-time digitisation, parsing, annotation and archival of eDNA data. The system tracks the provenance and analytical history of biological samples through a structured hierarchy of events, linked with associated digital file attachment archives, such as images and raw sequence files, and with inferred taxonomic occurrence records. The data entry process is compartmentalised and incorporated into the corresponding stages of standard operations used in fieldwork, biological collection management and laboratory analysis. Resulting data records can be integrated into various output formats required for large-scale analytics, publication and/or submission to global data aggregators. The prototype is implemented on the Microsoft 365 platform as a relational database (Access) linked to cloud-based data tables (SharePoint) and a set of associated data conversion spreadsheets (Excel). The system is designed primarily around the data management needs of small research labs; however, it is scalable to larger institutions and inter-institutional academic networks.

VisionDeep-AI: Deep learning-based retinal blood vessels segmentation and multi-class classification framework for eye diagnosis

Eye problems can lead to vision loss and have a significant impact on daily life, underscoring the critical importance of early diagnosis and treatment to prevent further damage and complications. A comprehensive segmentation-classification framework – VisionDeep-AI is developed in the proposed work for retinal vessel segmentation and multi-class classification on given fundus images. A weighted bi-directional feature pyramid network and U-Net backbone architecture-based customized deep-learning model is built to segment blood vessels which enhances feature extraction and multi-scale feature fusion. The architectural design comprises an end-to-end encoder-decoder network featuring six-depth layers with varying resolutions, enabling the extraction of high-level descriptors and lower-level, fine-grained characteristics. Furthermore, a multi-modal deep feature fusion architecture is developed for the multi-class classification of fundus images in four different categories by integrating features from segmented vessel images and raw fundus images, thereby accommodating more diversified information. A thorough analysis of VisionDeep-AI has been done on a comprehensive colour fundus image dataset. To ensure reliable results, data augmentation was done to prevent the over-fitting of the model and to enhance its generalization capability. The segmentation model performed exceptionally well and achieved a high accuracy of 97.73% and a dice coefficient of 89.90% in blood vessel segmentation from fundus images. Furthermore, multi-modal deep feature-fused classification architecture achieved a test accuracy of 81.50%, with a specificity of 93.83%. These findings showcase the robustness, generalizability and efficiency of the proposed VisionDeep-AI framework for fundus image segmentation and retinal disease classification, contributing towards advanced medical technologies and diagnosis precision.

A deep multi-task learning approach to identifying mummy berry infection sites, the disease stage, and severity.

Mummy berry is a serious disease that may result in up to 70 percent of yield loss for lowbush blueberries. Practical mummy berry disease detection, stage classification and severity estimation remain great challenges for computer vision-based approaches because images taken in lowbush blueberry fields are usually a mixture of different plant parts (leaves, bud, flowers and fruits) with a very complex background. Specifically, typical problems hindering this effort included data scarcity due to high manual labelling cost, tiny and low contrast disease features interfered and occluded by healthy plant parts, and over-complicated deep neural networks which made deployment of a predictive system difficult. Using real and raw blueberry field images, this research proposed a deep multi-task learning (MTL) approach to simultaneously accomplish three disease detection tasks: identification of infection sites, classification of disease stage, and severity estimation. By further incorporating novel superimposed attention mechanism modules and grouped convolutions to the deep neural network, enabled disease feature extraction from both channel and spatial perspectives, achieving better detection performance in open and complex environments, while having lower computational cost and faster convergence rate. Experimental results demonstrated that our approach achieved higher detection efficiency compared with the state-of-the-art deep learning models in terms of detection accuracy, while having three main advantages: 1) field images mixed with various types of lowbush blueberry plant organs under a complex background can be used for disease detection; 2) parameter sharing among different tasks greatly reduced the size of training samples and saved 60% training time than when the three tasks (data preparation, model development and exploration) were trained separately; and 3) only one-sixth of the network parameter size (23.98M vs. 138.36M) and one-fifteenth of the computational cost (1.13G vs. 15.48G FLOPs) were used when compared with the most popular Convolutional Neural Network VGG16. These features make our solution very promising for future mobile deployment such as a drone carried task unit for real-time field surveillance. As an automatic approach to fast disease diagnosis, it can be a useful technical tool to provide growers real time disease information that can prevent further disease transmission and more severe effects on yield due to fruit mummification.

Algorithm-Driven Tele-otoscope for Remote Care for Patients With Otitis Media.

The COVID-19 pandemic has spurred a growing demand for telemedicine. Artificial intelligence and image processing systems with wireless transmission functionalities can facilitate remote care for otitis media (OM). Accordingly, this study developed and validated an algorithm-driven tele-otoscope system equipped with Wi-Fi transmission and a cloud-based automatic OM diagnostic algorithm. Prospective, cross-sectional, diagnostic study. Tertiary Academic Medical Center. We designed a tele-otoscope (Otiscan, SyncVision Technology Corp) equipped with digital imaging and processing modules, Wi-Fi transmission capabilities, and an automatic OM diagnostic algorithm. A total of 1137 otoscopic images, comprising 987 images of normal cases and 150 images of cases of acute OM and OM with effusion, were used as the dataset for image classification. Two convolutional neural network models, trained using our dataset, were used for raw image segmentation and OM classification. The tele-otoscope delivered images with a resolution of 1280 × 720 pixels. Our tele-otoscope effectively differentiated OM from normal images, achieving a classification accuracy rate of up to 94% (sensitivity, 80%; specificity, 96%). Our study demonstrated that the developed tele-otoscope has acceptable accuracy in diagnosing OM. This system can assist health care professionals in early detection and continuous remote monitoring, thus mitigating the consequences of OM.

Intelligent Identification of Chinese and Australian Merino Wool Fibers Based on Image Recognition

In order to promote the sustainable growth of the wool industry and protect consumers' legitimate rights, rapid identification of the country of origin for wool of the same type is deemed crucial. This research presents a computer graphic recognition training model that utilizes median and Wiener filtering techniques to effectively reduce noise in the raw wool fiber images. Employing a support vector machine as the classifier and integrating a polynomial kernel function, this model achieves swift and accurate identification of Chinese and Australian Merino wool fibers. Experimental results underscore that following image recognition training, the model attains an impressive 92.5% comprehensive and precise identification rate for Chinese and Australian Merino wool fibers, effectively distinguishing the origin of wool from the same category. This approach not only provides a valuable reference for identifying the origin of similar wool types but also holds the potential to standardizing the wool fibre material market and assuring the consumer’s confidence on wool products.

NIEND: neuronal image enhancement through noise disentanglement.

The full automation of digital neuronal reconstruction from light microscopic images has long been impeded by noisy neuronal images. Previous endeavors to improve image quality can hardly get a good compromise between robustness and computational efficiency. We present the image enhancement pipeline named Neuronal Image Enhancement through Noise Disentanglement (NIEND). Through extensive benchmarking on 863 mouse neuronal images with manually annotated gold standards, NIEND achieves remarkable improvements in image quality such as signal-background contrast (40-fold) and background uniformity (10-fold), compared to raw images. Furthermore, automatic reconstructions on NIEND-enhanced images have shown significant improvements compared to both raw images and images enhanced using other methods. Specifically, the average F1 score of NIEND-enhanced reconstructions is 0.88, surpassing the original 0.78 and the second-ranking method, which achieved 0.84. Up to 52% of reconstructions from NIEND-enhanced images outperform all other four methods in F1 scores. In addition, NIEND requires only 1.6 s on average for processing 256 × 256 × 256-sized images, and images after NIEND attain a substantial average compression rate of 1% by LZMA. NIEND improves image quality and neuron reconstruction, providing potential for significant advancements in automated neuron morphology reconstruction of petascale. The study is conducted based on Vaa3D and Python 3.10. Vaa3D is available on GitHub (https://github.com/Vaa3D). The proposed NIEND method is implemented in Python, and hosted on GitHub along with the testing code and data (https://github.com/zzhmark/NIEND). The raw neuronal images of mouse brains can be found at the BICCN's Brain Image Library (BIL) (https://www.brainimagelibrary.org). The detailed list and associated meta information are summarized in Supplementary Table S3.

Study on the Ratio of Dependence of Non-Productive Age Population to Productive Age Population in West Nusa Tenggara, Indonesia

This paper aims to analyze and examine the level of dependence of the non-productive age population on the productive age population. There are two age categories of non-productive age population, namely the age of 0-14 years called the young age group, and the age category of 65 years and over. This study is a quantitative descriptive research, using secondary data in the form of raw data about productive age population and non-productive age population. Data analysis using the dependency ratio formula in total, young age, and old age. Data sources from the Central Bureau of Statistics and related institutions. The research was conducted in West Nusa Tenggara Province for the 2013-2022 time period (according to data availability). The results of the analysis / study showed that in total for the initial seven years, the burden of dependence was in the category of moderate dependence (85 percent of the total DR was the burden of youth), while for the last three years it has been in the low dependency category. Partially young age, indicating the burden of dependents is in the low category, as well as for old age is in the category of low dependent burden.

DiffRAW: Leveraging Diffusion Model to Generate DSLR-Comparable Perceptual Quality sRGB from Smartphone RAW Images

Deriving DSLR-quality sRGB images from smartphone RAW images has become a compelling challenge due to discernible detail disparity, color mapping instability, and spatial misalignment in RAW-sRGB data pairs. We present DiffRAW, a novel method that incorporates the diffusion model for the first time in learning RAW-to-sRGB mappings. By leveraging the diffusion model, our approach effectively learns the high-quality detail distribution of DSLR images, thereby enhancing the details of output images. Simultaneously, we use the RAW image as a diffusion condition to maintain image structure information such as contours and textures. To mitigate the interference caused by the color and spatial misalignment in training data pairs, we embed a color-position preserving condition within DiffRAW, ensuring that the output images do not exhibit color biases and pixel shift issues. To accelerate the inference process of DiffRAW, we designed the Domain Transform Diffusion Method, an efficient diffusion process with its corresponding reverse process. The Domain Transform Diffusion Method can reduce the required inference steps for diffusion model-based image restoration/enhancement algorithms while enhancing the quality of the generated images. Through evaluations on the ZRR dataset, DiffRAW consistently demonstrates state-of-the-art performance across all perceptual quality metrics (e.g., LPIPS, FID, MUSIQ), while achieving comparable results in PSNR and SSIM.