Histogram Bins Research Articles

Reversible Data Hiding With Brightness Preserving Contrast Enhancement by Two-Dimensional Histogram Modification

Recently, contrast enhancement with reversible data hiding (CE-RDH) has been proposed for digital images to hide useful data into contrast-enhanced images. In existing schemes, one-dimensional (1D) or two-dimensional (2D) histogram is equalized during the process of CE-RDH so that an original image can be exactly recovered from its contrast-enhanced version. However, noticeable brightness change and color distortion may be introduced by applying these schemes, especially in the case of over enhancement. To preserve image quality, this paper presents a new 2D histogram based CE-RDH scheme by taking brightness preservation (BP) into account. In particular, the row or column of histogram bins with the maximum total height are chosen to be expanded at each time of histogram modification, while the row or column of bins to be expanded next are adaptively chosen according to the change of image brightness. Experimental results on three color image sets demonstrate efficacy and reversibility of the proposed scheme. Compared with the schemes using 1D histogram, image brightness can be preserved more finely by modifying the generated 2D histogram. Moreover, our proposed scheme preserves image color and brightness while achieving better image quality than the existing schemes.

Lower rate limit in cardiac resynchronization therapy defibrillators

Abstract Background There is few data about programmed lower rate limit (LRL) in real world heart failure (HF) patients with cardiac resynchronization therapy–defibrillators (CRT-Ds) and its influence in clinical outcomes. Heart rate score (HRS) is the percentage of all atrial-paced and sensed events in the single tallest 10 beats/min device histogram bin and may indicate impaired heart rate variability. Purpose We hypothesized that higher LRL programming is associated with worse clinical outcomes as arrhythmic events and HF decompensations in chronic HF patients with CRT-Ds. Methods LRL was evaluated and HRS was calculated from remote monitoring in 126 HF patients with CRT-D. Primary outcome was defined as HF hospitalizations and related admissions to the emergency department and secondary outcome as number of device therapies, sustained ventricular tachycardia (VT) and ventricular fibrillation (VF). Results Mean age was 69,03±10,39 years, 81 (64,3%) were males and mean follow-up was 53,72±46,13 months. Mean left ventricular ejection fraction was 30,31±8,33% and 29 (23,0%) were in NYHA III–IV. HF aetiology was idiopathic in 39 (43,3%), ischemic in 32 (25,4%) and alcoholic cardiomyopathy in 8 (6,3%). Thirty-seven (29,4%) patients had atrial fibrillation and 33 (26,2%) coronary disease. LRL ranged from to 40 to 80 bpm and mean LRL was 52,64±9,64 and mean HRS 49,60±23,17%. Programmed LRL was higher in women (p=0,014), patients with atrial fibrillation (AF) (p=0,012) and coronary disease (p=0,015). Higher LRL correlated with HF hospitalizations and related admissions to the emergency department (ED) (r=0,541, p=0,001), VT or VF episodes (r=0,337, p=0,005) and CRT-D number of therapies (r=0,342, p=0,004) and higher HRS (r=0,547, p<0,05). Conclusion Higher LRL programming was associated with higher HRS, HF decompensations with hospitalization or admission to the emergency department, VT or VF episodes and CRT-D therapies in a real world population. More studies are required but lower LRL may be preferred in HF patients. Funding Acknowledgement Type of funding sources: None.

Robust Template Matching Using Multiple-Layered Absent Color Indexing.

Color is an essential feature in histogram-based matching. This can be extracted as statistical data during the comparison process. Although the applicability of color features in histogram-based techniques has been proven, position information is lacking during the matching process. We present a conceptually simple and effective method called multiple-layered absent color indexing (ABC-ML) for template matching. Apparent and absent color histograms are obtained from the original color histogram, where the absent colors belong to low-frequency or vacant bins. To determine the color range of compared images, we propose a total color space (TCS) that can determine the operating range of the histogram bins. Furthermore, we invert the absent colors to obtain the properties of these colors using threshold . Then, we compute the similarity using the intersection. A multiple-layered structure is proposed against the shift issue in histogram-based approaches. Each layer is constructed using the isotonic principle. Thus, absent color indexing and multiple-layered structure are combined to solve the precision problem. Our experiments on real-world images and open data demonstrated that they have produced state-of-the-art results. Moreover, they retained the histogram merits of robustness in cases of deformation and scaling.

Determining minimum sample size for the conditioned Latin hypercube sampling algorithm

In digital soil mapping (DSM), a fundamental assumption is that the spatial variability of the target variable can be explained by the predictors or environmental covariates. Strategies to adequately sample the predictors have been well documented, with the conditioned Latin hypercube sampling (cLHS) algorithm receiving the most attention in the DSM community. Despite advances in sampling design, a critical gap remains in determining the number of samples required for DSM projects. We propose a simple workflow and function coded in R language to determine the minimum sample size for the cLHS algorithm based on histograms of the predictor variables using the Freedman-Diaconis rule for determining optimal bin width. Data preprocessing was included to correct for multimodal and non-normally distributed data, as these can affect sample size determination from the histogram. Based on a user-selected quantile range (QR) for the sample plan, the densities of the histogram bins at the upper and lower bounds of the QR were used as a scaling factor to determine minimum sample size. This technique was applied to a field-scale set of environmental covariates for a well-sampled agricultural study site near Guelph, Ontario, Canada, and tested across a range of QRs. The results showed increasing minimum sample size with an increase in the QR selected. Minimum sample size increased from 44 to 83 when the QR increased from 50% to 95% and then increased exponentially to 194 for the 99% QR. This technique provides an estimate of minimum sample size that can be used as an input to the cLHS algorithm.

Do Borrowers Intentionally Avoid Covenant Violations? A Reexamination of the Debt Covenant Hypothesis

ABSTRACTIn this study, we replicate and extend the Dichev and Skinner [DS: 2002] study on the debt covenant hypothesis (DCH). We start by replicating DS and find results consistent with theirs. We then extend their work by changing three aspects of the research design: histogram bin width, calculation of slack, and statistical test of discontinuity. We find that the inference from DS is generally robust to varying these choices, although sensitive to different bin widths, during their sample period. We extend our analysis to the period 2000–2019 and find that support for DCH remains robust. We do, however, find a lack of support for DCH when examining the most common financial covenant, debt‐to‐EBITDA. These findings suggest a more nuanced perspective on DCH, whereby different types of financial covenants provide different incentives and abilities to avoid technical default.

Detection and Classification of Road Damage Based on Image Morphology and K-NN Method (K Nearest Neighbour)

Road pavement is a supporting factor for national development, especially in the distribution of trade in goods and services as well as the movement of human mobility. Road maintenance needs to be done regularly so that the road is always in good condition, but the weather and road loads are the things that cause road damage. Road damage is generally categorized into cracks, alligator cracks and potholes. The purpose of this research is to utilize image processing to detect and classify the types of road damage. The steps involved include: image acquisition with a digital camera, conversion of RGB images into grayscale images, image normalization, selection of damage points, counting histogram bins, determining damage bins, calculating noise with image morphology (closing and opening) using a disk element structure of size 5, calculating radial vector and finally classifying road damage using the K-NN (K Nearest Neighbor) method with 3 classes and a K value of 11. The image from the classification results is then calculated the level of damage based on the category according to the SDI (Surface Distress Index) provisions, where the level of crack damage is seen from the width of the crack, the alligator crack is seen from the percentage of damaged area compared to the segment under review and the pathole is seen from many holes. The test used 597 images consisting of 95% training data and 5% test data, the results obtained that the accuracy of this research reached 83%.

A novel severity calibration algorithm for defect detection by constructing maps

In industrial scenarios, knowledge of the severity of defects is needed to distinguish different levels of defects and adapt to the possible modification of defect identification criteria. Well-calibrated detection models are crucial, especially for these decision-making applications. Previous works have measured and calibrated biased confidence estimates of classification and object detection models, while the field of severity calibration has not yet been addressed. Through experiments, we find that the severity also has a specific relationship with the probabilistic prediction of models. Therefore, we present a framework to measure and calibrate severity calibration for defect detection applications by constructing maps between probabilistic prediction, precision, and severity. We find that unlike confidence calibration, the reliability diagram of a well-calibrated model for severity calibration is funnel-shaped instead of linear-shaped. To approximate the funnel-shaped mapping function, we introduce a Gaussian distribution -based function with parameters that are fitted by marked data. Additionally, extending algorithms are proposed to adjust classical confidence calibration methods (e.g., histogram binning, isotonic regression, vector scaling, and temperature scaling) for severity calibration. By mapping the probabilistic prediction to precision, severity calibration is transformed to confidence calibration. The proposed severity calibration methods are evaluated by the expected severity calibration error (ESCE) in comparative experiments, which show that the extended isotonic regression achieves the best performance. In Zhongce’s 21 intelligent detection lines, we applied severity calibration and received good feedback from inspectors.

Stochasticity in threshold of homologous recombination deficiency (HRD) score.

e17561 Background: HRD test from Myriad Inc. was approved by FDA as a biomarker for niraparib in ovarian cancer. Its threshold (as 42) was initially established by capturing 95% of tumors with BRCA1/2 mutation and BRCA1 promoter methylation in the training cohort. Here, we explored the stochasticity of such threshold, when the test noises themselves were concerned. Methods: We simulated HRD scores from figure 1 of Telli et al., 2016 by assigning the median from each histogram bin to the associated samples. The threshold was calculated, as described in the original paper, and the standard error (SE) and confidence interval (CI) for such threshold were estimated from bootstrapping the cohort. Furthermore, the entire cohort was down-sampled to determine the effect of sample size on the threshold. Results: The threshold of 42.5 was calculated from our simulated dataset, close to the original one, with SE of 1.62 and 95% CI of 40.75 - 47.5 for the bootstrap. As expected, SE steadily increased to 2.21, 2.84, 2.97 when the sample sizes decreased to 250, 230, and 210 respectively. When the dataset was further down-sampled to include only 200 BRCA deficient samples, SE increased to 3.76, and drastically, CI ranged between 36.5 and 51.2. Conclusions: We have showed there exists manifest stochasticity in the original threshold of 42 for the Myriad HRD test, which is likely close to platform noise in magnitude. The minimum size of BRCA deficient patients to obtain a stable threshold is 200, as shown in our study.

Design and Analysis of Reversible Data Hiding Using Hybrid Cryptographic and Steganographic approaches for Multiple Images

Data concealing is the process of including some helpful information on images. The majority of sensitive applications, such sending authentication data, benefit from data hiding. Reversible data hiding (RDH), also known as invertible or lossless data hiding in the field of signal processing, has been the subject of a lot of study. A piece of data that may be recovered from an image to disclose the original image is inserted into the image during the RDH process to generate a watermarked image. Lossless data hiding is being investigated as a strong and popular way to protect copyright in many sensitive applications, such as law enforcement, medical diagnostics, and remote sensing. Visible and invisible watermarking are the two types of watermarking algorithms. The watermark must be bold and clearly apparent in order to be visible. To be utilized for invisible watermarking, the watermark must be robust and visibly transparent. Reversible data hiding (RDH) creates a marked signal by encoding a piece of data into the host signal. Once the embedded data has been recovered, the original signal may be accurately retrieved. For photos shot in poor illumination, visual quality is more important than a high PSNR number. The DH method increases the contrast of the host picture while maintaining a high PSNR value. Histogram equalization may also be done concurrently by repeating the embedding process in order to relocate the top two bins in the input image's histogram for data embedding. It's critical to assess the images after data concealment to see how much the contrast has increased. Common picture quality assessments include peak signal to noise ratio (PSNR), relative structural similarity (RSS), relative mean brightness error (RMBE), relative entropy error (REE), relative contrast error (RCE), and global contrast factor (GCF). The main objective of this paper is to investigate the various quantitative metrics for evaluating contrast enhancement. The results show that the visual quality may be preserved by including a sufficient number of message bits in the input photographs.

Labeling of an intra-class variation object in deep learning classification

<span lang="EN-US">Machine orientation learning had demonstrated that deep learning (DL)-convolutional neural networks (CNNs) were robust image classifiers with significant accuracy. Although to been functional, DL scope classification as tight, well-defined as possible uses a 2-class object, for instance, cats and dogs. The DL classification faced many challenges, e.g., variation factors, the intra-class variation. This nature is presented in every object, its diversity of an object. The label was an exact given name of an intra-class variation object. Unfortunately, not every object had a specific name, in exceptionally high similarity inside the category. This paper explored those problems in flower plants’ taxonomy naming. In supervised learned of DL, image datasets musted labeled with a meaningful word or phrase that humans are familiar with, a taxonomy naming. Labeled with visual feature extraction brought a fully automatic classification. Flower Plumeria L labeling extracted from perspective dimension scale of petal flower which automatically obtained by contour detection, and peaks of blue green red (BGR) histogram channels from bins histogram after object masked. Dataset collected on photography workbench equipped with webcam and ring light. Results showed labels for intra-class variation of Plumeria L in form of dimension-scale and BGR-peaks. The result of this study presented a novelty in building datasets for intra-class variation for the DL classification.</span>

A Supervoxel Segmentation Method With Adaptive Centroid Initialization for Point Clouds

Supervoxels find applications as a pre-processing step in many image processing problems due to their ability to present a regional representation of points by correlating them into a set of clusters. Besides reducing the overall computational time for subsequent algorithms, the desirable properties in supervoxels are adherence to object boundaries and compactness. Existing supervoxel segmentation methods define the size of a supervoxel based on a user inputted resolution value. A fixed resolution results in poor performance in point clouds with non-uniform density. Whereas, other methods, in their quest for better boundary adherence, produce supervoxels with irregular shapes and elongated boundaries. In this article, we propose a new supervoxel segmentation method, based on k-means algorithm, with dynamic cluster seed initialization to ensure uniform distribution of cluster seeds in point clouds with variable densities. We also propose a new cluster seed initialization strategy, based on histogram binning of surface normals, for better boundary adherence. Our algorithm is parameter-free and gives equal importance to the color, spatial location and orientation of the points resulting in compact supervoxels with tight boundaries. We test the efficacy of our algorithm on a publicly available point cloud dataset consisting of 1449 pairs of indoor RGB-D images, i.e., color (RGB) images coupled with depth information (D) mapped per pixel. Results are compared against three state-of-the-art algorithms based on four quality metrics. Results show that our method provides significant improvement over other methods in the undersegmentation error and compactness metrics and, performs equally well in the boundary recall and contour density metrics.

A study of the effects of histogram binning on the accuracy of finding flux distribution of X-ray binaries

X-ray binaries consist of a normal star and a collapsed star which could be either a white dwarf, neutron star, or a black hole. X-ray emission from them exhibits strong variability at time scales from milliseconds to years. The variation in flux across the electromagnetic spectrum over a wide range of time scales is an extensive characteristic of an astrophysical body powered by a compact source. Flux distribution is an important tool to understand the variability processes in X-ray binaries. A normal flux distribution suggests additive processes while a log normal refers to cascade/multiplicative processes. Histogram fitting is one of the statistical methods which could be implemented to study this distribution. In our study we have observed that with binning with large number of data points gives a better results in comparison to a small number of data points.

Lower rate limit for pacing by cardiac resynchronization defibrillators: Should lower rate programming be reconsidered?

No real-world large database associates lower rate limit (LRL) programming and survival of subjects with cardiac resynchronization therapy-defibrillators (CRT-Ds). The purpose of this study was to test the hypothesis that lower LRL programming is independently associated with survival, and that LRL and heart rate score (HrSc) are associated. All dual-chamber CRT-D devices in the Remote Patient Monitoring (RPM) ALTITUDE database (2006-2011) were queried. Baseline HrSc was defined as the percentage of all atrial sensed and paced beats in the tallest 10-beat histogram bin postimplant. LRL was assessed during repeated RPM uploads. Using a Cox model multivariable analysis, relationships between LRL, survival, HrSc, and other variables were evaluated. Survival was determined by query of death indices. Data analyzed included 61,881 subjects (mean follow-up 2.9 years). LRL ranged from 40 to 85 bpm. Baseline lower LRL was associated with younger age, less atrial fibrillation, female sex, and lower HrSc (P <.001 for all covariates). Lower LRL was associated with improved survival, with LRL 40 associated with the largest survival benefit. This was significant for all 3 HrSc subgroups (P<.001). An interaction between HrSc and LRL was observed, with the largest survival difference between HrSc groups observed at LRL-40 (P <.001). LRL programming and HrSc were associated, and lower values of both were associated with improved survival in a large database of CRT-D subjects. Relationships between survival, LRL programming, and HrSc merit further study.

Scale-invariant histogram of oriented gradients: novel approach for pedestrian detection in multiresolution image dataset

This paper proposes a scale-invariant histogram of oriented gradients (SI-HOG) for pedestrian detection. Most of the algorithms for pedestrian detection use the HOG as the basic feature and combine other features with the HOG to form the feature set, which is usually applied with a support vector machine (SVM). Hence, the HOG feature is the most efficient and fundamental feature for pedestrian detection. However, the HOG feature produces feature vectors of different lengths for different image resolutions; thus, the feature vectors are incomparable for the SVM. The proposed method forms a scale-space pyramid wherein the histogram bin is calculated. Thus, the gradient information from all the scales is encapsulated in a single fixed-length feature vector. The proposed method is also combined with color and texture features. The proposed approach is tested on three established benchmark pedestrian datasets: INRIA, NICTA, and Daimler. An improvement of ≥4.5% in the miss rate is achieved for all the three datasets considered. We also show that the SI-HOG can be applied to multiresolution datasets for which the HOG feature cannot be applied. Additionally, the MapReduce model is used to obtain the same outcome. The results indicate that the proposed approach outperforms the pedestrian-detection methods considered in this work.

Heart rate score and outcomes in ICD patients: insights from the prospective randomized INTRINSIC RV trial

Heart rate score (HRSc), the percentage of atrial sensed and paced beats in the largest 10 beat/min bin of a device histogram and mean intrinsic heart rate (MIHR), predicted survival in nonrandomized studies of implantable defibrillator (ICD) patients. We evaluated whether HRSc and MIHR independently predicted mortality and heart failure (HF) hospitalization in the prospective, randomized, controlled INTRINSIC RV trial. The INTRINSIC RV trial enrolled 1530 patients receiving dual-chamber ICDs. This analysis involves patients (n = 1471) for whom MIHR and HRSc data were available. Mean follow-up was 10.4months. The relationship between pre-randomization MIHR and HRSc on the primary endpoint of all-cause mortality and HF hospitalization was assessed using multivariate regression and Cox modeling. As categorical variables, MIHR > 70bpm and HRSc > 70% were considered high. The median baseline MIHR and HRSc were 74 (IQR = 16) and 50% (IQR = 20) respectively. As a continuous variable, for every 1% increase in HRSc, death/HF hospitalization increased by 1% (95%CI: 1.002-1.017; p = 0.01). Regression analysis showed baseline MIHR was associated with HRSc (p = 0.01); for every 1 beat/min increase in MIHR, HRSc increased by 1.8%. A MIHR > 70bpm and HRSc ≥ 70% predicted, but were independently associated with, the primary endpoint (HR: 1.39; 95%CI: 1.10-1.76, p = 0.005 for MIHR and HR: 1.654; 95%CI: 1.11-2.46, p = 0.01 for HRSc). Male gender (HR: 0.75), history of HF (HR: 1.29), and atrial fibrillation (HR: 1.37) also predicted death/hospitalization in the Cox model. In this large, prospectively studied ICD population, HRSc was a robust and independent predictor of death/HF hospitalization. High MIHR and high HRSc were associated but each predicted outcomes independently.

Photographic Reproduction and Enhancement Using HVS-Based Modified Histogram Equalization.

Photographic reproduction and enhancement is challenging because it requires the preservation of all the visual information during the compression of the dynamic range of the input image. This paper presents a cascaded-architecture-type reproduction method that can simultaneously enhance local details and retain the naturalness of original global contrast. In the pre-processing stage, in addition to using a multiscale detail injection scheme to enhance the local details, the Stevens effect is considered for adapting different luminance levels and normally compressing the global feature. We propose a modified histogram equalization method in the reproduction stage, where individual histogram bin widths are first adjusted according to the property of overall image content. In addition, the human visual system (HVS) is considered so that a luminance-aware threshold can be used to control the maximum permissible width of each bin. Then, the global tone is modified by performing histogram equalization on the output modified histogram. Experimental results indicate that the proposed method can outperform the five state-of-the-art methods in terms of visual comparisons and several objective image quality evaluations.

Investigation of volume fraction of GaP nanowires by SEM characterization and spectroscopic ellipsometry

In this paper we focus on investigation of GaP volume fraction in GaP nanowires (NWs) prepared on GaP substrates by MOVPE using VLS mode. For this purpose, spectroscopic ellipsometry (SE) and scanning electron microscopy (SEM) image processing methods were used showing a good agreement. In ellipsometry, composition of GaP and void was modeled using linear gradient model in combination with Bruggeman Effective Medium Approximation (BEMA). From this model the volume fraction of GaP on the top of the nanowires was determined from 1.5%–6.3%) and in the bottom of the nanowires from 34.8%–39.6%), respectively. The histogram of GaP nanowires SEM image was divided into 10 bins above the plane. The volume fraction of GaP nanowires determined from these bins ranged from 1.4 % to 3.8 % on the top of the nanowires and from 33.5%–46.2% in their bottom. Linearity of the SEM histogram bins gradient was further compared to the linearity of the linear gradient layer model.

Heart rate score, a measure related to chronotropic incompetence in pacemaker patients.

BackgroundHeart rate score (HrSc) ≥70% in cardiac resynchronization therapy defibrillator and implantable cardioverter-defibrillator subjects predicts 5-year mortality risk. A high HrSc suggests few sensed cardiac cycles above the programmed lower rate.ObjectiveTo determine if HrSc is related to chronotropic incompetence (CI) in pacemaker (PM) subjects.MethodsHrSc is the percentage of all atrial-paced and sensed events in the single tallest 10 beats/min histogram bin programmed to DDD 60/min. The prospective LIFE study of PM subjects examined multiple treadmill-based measures of CI. The 1-month postimplant DDD 60/min PM rate histogram prior to treadmill was retrospectively analyzed for HrSc. Measures of CI were applied to submaximal treadmill data in the DDD mode. HrSc was compared to these CI measures and to clinical indications for PM.ResultsThe 1-month histogram demonstrated HrSc ≥70% in 43% of subjects. HrSc ≥70% correlated with a clinical diagnosis of sick sinus syndrome (P < .001). CI was present in 34%–88% of subjects by treadmill-based measures. Agreement between treadmill-based measures for CI was poor and varied from 39% to 83%. HrSc ≥70%, as a measure of CI, was most highly correlated with unpaced heart rate <70% of age-predicted maximum heart rate (67%) (odds ratio 3.7, P < .001).ConclusionsHrSc ≥70% correlates with treadmill measures of CI and clinical sick sinus syndrome. HrSc ≥70% is a measure of CI in PM subjects that is inexpensive, repeatable, and quantitative.

The perfusion index histograms predict patent ductus arteriosus requiring treatment in preterm infants.

The impact of patent ductus arteriosus (PDA) on vital sign trends represented as histograms, and perfusion index in particular, is unknown. This study aimed to split continuously obtained PI and other vital signs before, during, and after medical treatment of PDA, into histogram bins, and determine the utility of PI and other vital sign histograms in the early prediction of hemodynamically significant PDA (hsPDA). In 34 infants at a mean gestational age of 26 ± 2.1weeks, we prospectively collected vital signs for three different periods, 24h before starting treatment of PDA, during PDA treatment, and 24h after completion of the course of treatment, and confirmed PDA closure by echo. Histograms with three comparable periods were obtained from preterm infants who did not require treatment for PDA and analyzed for comparison. The duration of time spent in each histogram bin was determined for each time epoch. Episodes of low PI < 0.4 and high PI > 2 were significantly longer in duration in infants with PDA before treatment compared to those in infants with PDA during and after treatment. The arterial oxygen saturation (SpO2) < 80% was also longer in duration in infants with PDA before compared to that in infants with PDA during and after treatment. Low PI < 0.4 correlated with most echocardiography indices of hsPDA.Conclusion: We conclude that a patent ductus arteriosus requiring treatment in preterm infants ≤ 29weeks GA was associated with significant fluctuations between a low PI < 0.4 alternating with a high PI > 2, reflecting the dynamic nature of hsPDA shunt volume. PI variability may be an early marker of hsPDA. What is Known: • The perfusion index is a continuous underutilized parameter provided by pulse oximetry to assess the peripheral perfusion. • The perfusion index helps predict conditions with hemodynamic instability. What is New: • The perfusion index assessed as daily histogram trends can predict patent ductus arteriosus requiring treatment.

A Novel Method for Quantitative Analysis of C-Reactive Protein Lateral Flow Immunoassays Images via CMOS Sensor and Recurrent Neural Networks.

Objective: To design and implement an easy-to-use, Point-of-Care (PoC) lateral flow immunoassays (LFA) reader and data analysis system, which provides a more in-depth quantitative analysis for LFA images than conventional approaches thereby supporting efficient decision making for potential early risk assessment of cardiovascular disease (CVD). Methods and procedures: A novel end-to-end system was developed including a portable device with CMOS camera integrated with optimized illumination and optics to capture the LFA images produced using high-sensitivity C-Reactive Protein (hsCRP) (concentration level < 5 mg/L). The images were transmitted via WiFi to a back-end server system for image analysis and classification. Unlike common image classification approaches which are based on averaging image intensity from a region-of-interest (ROI), a novel approach was developed which considered the signal along the sample’s flow direction as a time series and, consequently, no need for ROI detection. Long Short-Term Memory (LSTM) networks were deployed for multilevel classification. The features based on Dynamic Time Warping (DTW) and histogram bin counts (HBC) were explored for classification. Results: For the classification of hsCRP, the LSTM outperformed the traditional machine learning classifiers with or without DTW and HBC features performed the best (with mean accuracy of 94%) compared to other features. Application of the proposed method to human plasma also suggests that HBC features from LFA time series performed better than the mean from ROI and raw LFA data. Conclusion: As a proof of concept, the results demonstrate the capability of the proposed framework for quantitative analysis of LFA images and suggest the potential for early risk assessment of CVD. Clinical impact: The hsCRP levels < 5 mg/L were aligned with clinically actionable categories for early risk assessment of CVD. The outcomes demonstrated the real-world applicability of the proposed system for quantitative analysis of LFA images, which is potentially useful for more LFA applications beyond presented in this study.