Automatic Background Research Articles

Automatic hierarchical background virtualization method for monocular vision image based on depth information extraction

Automatic hierarchical background virtualization method for monocular vision image based on depth information extraction

Automatic detection of cardiac conditions from photos of electrocardiogram captured by smartphones

BackgroundResearchers have developed machine learning-based ECG diagnostic algorithms that match or even surpass cardiologist level of performance. However, most of them cannot be used in real-world, as older generation ECG...

ProteinCoLoc streamlines Bayesian analysis of colocalization in microscopic images

Colocalization, the spatial overlap of molecular entities, is often key to support their involvement in common functions. Existing colocalization tools, however, face limitations, particularly because of their basic statistical analysis and their low-throughput manual entry processes making them unsuitable for automation and potentially introducing bias. These shortcomings underscore the need for user-friendly tools streamlining colocalization assessments and enabling their robust and automated quantitative analyses. We have developed ProteinCoLoc, an innovative software designed for automated high-throughput colocalization analyses and incorporating advanced statistical features such as Bayesian modelling, automatic background detection and localised correlation analysis. ProteinCoLoc rationalises colocalization assessments without manual input, comes with a user-friendly graphical user interface and provides various analytics allowing to study and locally quantify colocalization. This easy-to-use application presents numerous advantages, including a direct comparison with controls employing a Bayesian model and the analysis of local correlation patterns, while reducing hands-on time through automatic background detection. The software was validated while studying the colocalization pattern of two proteins forming a stable complex: the huntingtin protein (HTT) and its partner huntingtin-associated protein 40 (HAP40). Our results showcase the software’s capacity to quantitatively assess colocalizations. ProteinCoLoc is available both as a Julia package and as a compiled software (https://github.com/ma-seefelder/ProteinCoLoc).

Gamma-Ray Burst Detection with Poisson-FOCuS and Other Trigger Algorithms

We describe how a novel online change-point detection algorithm, called Poisson-FOCuS, can be used to optimally detect gamma-ray bursts within the computational constraints imposed by miniaturized satellites such as the upcoming HERMES-Pathfinder constellation. Poisson-FOCuS enables testing for gamma-ray burst onset at all intervals in a count time series, across all timescales and offsets, in real time and at a fraction of the computational cost of conventional strategies. We validate an implementation with automatic background assessment through exponential smoothing, using archival data from Fermi-GBM. Through simulations of lightcurves modeled after real short and long gamma-ray bursts, we demonstrate that the same implementation has higher detection power than algorithms designed to emulate the logic of Fermi-GBM and Compton-BATSE, reaching the performance of a brute-force benchmark with oracle information on the true background rate, when not hindered by automatic background assessment. Finally, using simulated data with different lengths and means, we show that Poisson-FOCuS can analyze data twice as fast as a similarly implemented benchmark emulator for the historic Fermi-GBM on-board trigger algorithms.

Multiplex, Quantitative, High-Resolution Imaging of Protein:Protein Complexes via Hybridization Chain Reaction.

Signal amplification based on the mechanism of hybridization chain reaction (HCR) facilitates spatial exploration of gene regulatory networks by enabling multiplex, quantitative, high-resolution imaging of RNA and protein targets. Here, we extend these capabilities to the imaging of protein:protein complexes, using proximity-dependent cooperative probes to conditionally generate a single amplified signal if and only if two target proteins are colocalized within the sample. HCR probes and amplifiers combine to provide automatic background suppression throughout the protocol, ensuring that even if reagents bind nonspecifically in the sample, they will not generate amplified background. We demonstrate protein:protein imaging with a high signal-to-background ratio in human cells, mouse proT cells, and highly autofluorescent formalin-fixed paraffin-embedded (FFPE) human breast tissue sections. Further, we demonstrate multiplex imaging of three different protein:protein complexes simultaneously and validate that HCR enables accurate and precise relative quantitation of protein:protein complexes with subcellular resolution in an anatomical context. Moreover, we establish a unified framework for simultaneous multiplex, quantitative, high-resolution imaging of RNA, protein, and protein:protein targets, with one-step, isothermal, enzyme-free HCR signal amplification performed for all target classes simultaneously.

Automatic Depth Estimation and Background Blurring of Animated Scenes Based on Deep Learning

Automatic Depth Estimation and Background Blurring of Animated Scenes Based on Deep Learning

Time frequency domain deep CNN for automatic background classification in speech signals

Many application areas, such as background identification, predictive maintenance in industrial applications, smart home applications, assisting deaf people with their daily activities and indexing and retrieval of content-based multimedia, etc., use automatic background classification using speech signals. It is challenging to predict the background environment accurately from speech signal information. Thus, a novel synchrosqueezed wavelet transform (SWT)-based deep learning (DL) approach is proposed in this paper for automatically classifying background information embedded in speech signals. Here, SWT is incorporated to obtain the time-frequency plot from the speech signals. These time-frequency signals are then fed to a deep convolutional neural network (DCNN) to classify background information embedded in speech signals. The proposed DCNN model consists of three convolution layers, one batch-normalization layer, three max-pooling layers, one dropout layer, and one fully connected layer. The proposed method is tested using various background signals embedded in speech signals, such as airport, airplane, drone, street, babble, car, helicopter, exhibition, station, restaurant, and train sounds. According to the results, the proposed SWT-based DCNN approach has an overall classification accuracy of 97.96 (± 0.53)% to classify background information embedded in speech signals. Finally, the performance of the proposed approach is compared to the existing methods.

Automatic background correction method for laser-induced breakdown spectroscopy

In laser-induced breakdown spectroscopy (LIBS), fluctuations in laser energy, fluctuations in laser-sample or laser-plasma interactions, and environmental noise cause the acquired spectral to contain different backgrounds. The background has a substantial impact on the analysis of spectra. The aim of this work is to present a method for automatic estimation of diverse spectral backgrounds. The proposed method utilizes the ideas of window functions and differentiation, combined with a piecewise cubic Hermite interpolating polynomial (Pchip), to achieve automatic removal of spectral backgrounds. In simulation experiments on background correction, it has been observed that this method outperforms asymmetric least squares (ALS) and Model-free background correction in terms of effectiveness. It exhibits the highest signal-to-background ratio (SBR). This method effectively eliminates the baseline that makes the spectrum elevated and some of the white noise. It also demonstrates stability in the background baseline jumps and in the dense region of the characteristic spectral lines. By using this approach to correct the spectral backgrounds of seven different aluminum alloys, there is a notable enhancement in the linear correlation coefficient between spectral intensity and the concentration of Mg element. In the experiment of measuring Mg concentration in aluminum alloys,the linear correlation coefficients between the predicted and actual concentrations of Mg in aluminum alloys were substantially improved after correction of the spectral background. The values of linear correlation coefficients after background correction using asymmetric least squares (ALS) and model-free methods were 0.9913 and 0.9926, respectively. The correlation coefficients between predicted and actual concentrations of Mg in aluminum alloys were improved from 0.9154 to 0.9943 by using the method proposed in this study. In contrast to ALS and Model-free methods, the proposed method showed more significant improvement on this experiment with higher linear correlation and smaller error. Both simulation and quantitative experimental findings affirm that this method achieves high accuracy and can be effectively employed for background correction prior to quantitative analysis.

Antenna effect of perylene-sensitized up-conversion luminescent material amplifies the signal of electrochemiluminescence biosensor platform for the ultra-sensitive detection of enrofloxacin

Recently, up-conversion luminescent (UCL) materials have caught extensive sight on account of their excellent biocompatibility and weak automatic fluorescence background, but the low optical signal makes researchers shy away. Organic dye-sensitized UCL materials can improve the low optical signal drawback of UCL and rejuvenate it with adjustable optical properties and unique antenna effects. In this work, an efficient, simple and selective electrochemiluminescence (ECL) sensing platform was developed for determination of enrofloxacin (ENR). 3,4,9,10-perylene tetracarboxylic acid (PTCA) was successfully used as an “antenna” to improve the ECL performance of the UCL nanoparticles (PEI-NaYF4: Yb, Er) due to its appropriate excitation spectrum position and superior electron transfer rate. The specific recognition function of the aptamer enabled the sensor to eliminate the interference from conspecific impurity. In the presence of ENR, the specific combination of ENR with aptamer made the aptamer fall from surface of the electrode, thus we could see a considerable enhancement of signal. Under the most favourable conditions, the aptasensor based on antenna effect displayed a wide detection range (1.0 × 10−14∼1.0 × 10−6 M), low limit of detection (LOD = 3.0 × 10−15 M) and receivable recoveries (96.0%–102.4%) during water samples analysis. At this point, antenna effect provides a powerful strategy to expand the application of UCL in the field of ECL biosensing.

Knowledge-enhanced Prompt-tuning for Stance Detection

Investigating public attitudes on social media is important in opinion mining systems. Stance detection aims to analyze the attitude of an opinionated text (e.g., favor, neutral, or against) toward a given target. Existing methods mainly address this problem from the perspective of fine-tuning. Recently, prompt-tuning has achieved success in natural language processing tasks. However, conducting prompt-tuning methods for stance detection in real-world remains a challenge for several reasons: (1) The text form of stance detection is usually short and informal, which makes it difficult to design label words for the verbalizer. (2) The tweet text may not explicitly give the attitude. Instead, users may use various hashtags or background knowledge to express stance-aware perspectives. In this article, we first propose a prompt-tuning-based framework that performs stance detection in a cloze question manner. Specifically, a knowledge-enhanced prompt-tuning framework (KEprompt) method is designed, which consists of an automatic verbalizer (AutoV) and background knowledge injection (BKI). Specifically, in AutoV, we introduce a semantic graph to build a better mapping from the predicted word of the pretrained language model and detection labels. In BKI, we first propose a topic model for learning hashtag representation and introduce ConceptGraph as the supplement of the target. At last, we present a challenging dataset for stance detection, where all stance categories are expressed in an implicit manner. Extensive experiments on a large real-world dataset demonstrate the superiority of KEprompt over state-of-the-art methods.

Deep-Learning-Based Automated Tracking and Counting of Living Plankton in Natural Aquatic Environments.

Plankton are widely distributed in the aquatic environment and serve as an indicator of water quality. Monitoring the spatiotemporal variation in plankton is an efficient approach to forewarning environmental risks. However, conventional microscopy counting is time-consuming and laborious, hindering the application of plankton statistics for environmental monitoring. In this work, an automated video-oriented plankton tracking workflow (AVPTW) based on deep learning is proposed for continuous monitoring of living plankton abundance in aquatic environments. With automatic video acquisition, background calibration, detection, tracking, correction, and statistics, various types of moving zooplankton and phytoplankton were counted at a time scale. The accuracy of AVPTW was validated with conventional counting via microscopy. Since AVPTW is only sensitive to mobile plankton, the temperature- and wastewater-discharge-induced plankton population variations were monitored online, demonstrating the sensitivity of AVPTW to environmental changes. The robustness of AVPTW was also confirmed with natural water samples from a contaminated river and an uncontaminated lake. Notably, automated workflows are essential for generating large amounts of data, which are a prerequisite for available data set construction and subsequent data mining. Furthermore, data-driven approaches based on deep learning pave a novel way for long-term online environmental monitoring and elucidating the correlation underlying environmental indicators. This work provides a replicable paradigm to combine imaging devices with deep-learning algorithms for environmental monitoring.

Automatic Inhomogeneous Background Correction for Spatial Target Detection Image Based on Partition Processing

High-resolution imaging with wide field of view (FoV) ground-based telescopes is often affected by skylight background and noise due to the detector, resulting in an inhomogeneous background. In this paper, we propose an improved method for spatial image non-uniformity correction based on partition processing. First, an evaluation metric is introduced to evaluate the partition size and automatically iterate a suitable partition value for different scenarios based on the different operating conditions of the telescope. Then, we iteratively calculate the mean and variance in each partitioned region to extract the background of each partitioned region. Finally, after applying bilinear interpolation to the background extracted from each region, the inhomogeneous background is obtained and removed from the original image. The experiments on the simulated and real images show that the proposed method can effectively remove the inhomogeneous background of spatial images and meet the requirements of the real-time processing of high-resolution images under long exposure conditions.

A Novel OpenBCI Framework for EEG-Based Neurophysiological Experiments.

An Open Brain-Computer Interface (OpenBCI) provides unparalleled freedom and flexibility through open-source hardware and firmware at a low-cost implementation. It exploits robust hardware platforms and powerful software development kits to create customized drivers with advanced capabilities. Still, several restrictions may significantly reduce the performance of OpenBCI. These limitations include the need for more effective communication between computers and peripheral devices and more flexibility for fast settings under specific protocols for neurophysiological data. This paper describes a flexible and scalable OpenBCI framework for electroencephalographic (EEG) data experiments using the Cyton acquisition board with updated drivers to maximize the hardware benefits of ADS1299 platforms. The framework handles distributed computing tasks and supports multiple sampling rates, communication protocols, free electrode placement, and single marker synchronization. As a result, the OpenBCI system delivers real-time feedback and controlled execution of EEG-based clinical protocols for implementing the steps of neural recording, decoding, stimulation, and real-time analysis. In addition, the system incorporates automatic background configuration and user-friendly widgets for stimuli delivery. Motor imagery tests the closed-loop BCI designed to enable real-time streaming within the required latency and jitter ranges. Therefore, the presented framework offers a promising solution for tailored neurophysiological data processing.

Research on Simulation Technology of Photoelectric Theodolite Based on Measured Infrared Image Sequence

Infrared image simulation[1] is an effective means to test, train and evaluate photoelectric theodolite, which has very important economic and military value. To solve the problems of weak pertinence, low fidelity, and complex algorithm of traditional infrared image simulation technology, this paper proposes a new infrared simulation technology based on a measured image sequence. Through automatic background stitching, moving target detection and extraction, a real-time fusion of target background, and infrared image simulation of real target/background is realized. The experimental results show that the technique has high accuracy and good real-time performance, and provides reliable technical support for the performance evaluation and personnel training of photoelectric theodolite.

New machine learning-based automatic high-throughput video tracking system for assessing water toxicity using Daphnia Magna locomotory responses

Daphnia magna is an important organism in ecotoxicity studies because it is sensitive to toxic substances and easy to culture in laboratory conditions. Its locomotory responses as a biomarker are highlighted in many studies. Over the last several years, multiple high-throughput video tracking systems have been developed to measure the locomotory responses of Daphnia magna. These high-throughput systems, used for high-speed analysis of multiple organisms, are essential for efficiently testing ecotoxicity. However, existing systems are lacking in speed and accuracy. Specifically, speed is affected in the biomarker detection stage. This study aimed to develop a faster and better high-throughput video tracking system using machine learning methods. The video tracking system consisted of a constant temperature module, natural pseudo-light, multi-flow cell, and an imaging camera for recording videos. To measure Daphnia magna movements, we developed a tracking algorithm for automatic background subtraction using k-means clustering, Daphnia classification using machine learning methods (random forest and support vector machine), and tracking each Daphnia magna location using the simple online real-time tracking algorithm. The proposed tracking system with random forest performed the best in terms of identification (ID) precision, ID recall, ID F1 measure, and ID switches, with scores of 79.64%, 80.63%, 78.73%, and 16, respectively. Moreover, it was faster than existing tracking systems such as Lolitrack and Ctrax. We conducted an experiment to observe the impact of toxicants on behavioral responses. Toxicity was measured manually in the laboratory and automatically using the high-throughput video tracking system. The median effective concentration of Potassium dichromate measured in the laboratory and using the device was 1.519 and 1.414, respectively. Both measurements conformed to the guideline provided by the Environmental Protection Agency of the United States; therefore, our method can be used for water quality monitoring. Finally, we observed Daphnia magna behavioral responses in different concentrations after 0, 12, 18, and 24 h and found that there was a difference in movement according to the concentration at all hours.

Footwear segmentation and recommendation supported by deep learning: an exploratory proposal

The management of an online footwear retail store - also known as marketplace - usually involves activities that directly or indirectly end up to interface with customers, wherein communication efficiency and effectiveness is crucially relevant. Critical factors concerning entities developing remote business in these areas or similar include: (i) production of appealing catalogues and (ii) digital tools to shorten the distance between customers and marketplaces. The former requires using specific third parties - often technically complex - to arrange and prepare photographic entries acquired in studio-like environments. This can delay the diffusion of products supply that may result in financial losses. The latter prevents the retailer of reaching critical mass at a higher potential. Considering such issues, this paper proposes a couple of modules for footwear marketplaces, powered by deep learning: one to segment shoes as a fully automatic background removal tool for easing and quickening catalogue creation activities in a back-office perspective; and another to provide visual search services that allow a customer to submit photographs of footwear of interest to obtain recommendations of similar products directly retrieved from online retail databases, establishing another digital bridge with potential buyers. Preliminary implementation and pilot tests point out Mask-RCNN as a promising approach for shoes segmentation. The same applies to ResNet101 and Xception, but for shoes recommendation, based on multi-label classification.

Lock-In Amplifier for Sensor Application Using Second Order Harmonic Frequency With Automatic Background Phase Calibration

An integrated lock-in amplifier (LIA) with automatic phase tuning and second-order harmonic frequency extraction is presented in this paper. The automatic phase alignment loop implemented in the proposed single-channel LIA can directly align the phases of an input signal with a reference signal. Furthermore, the proposed LIA system also includes an enhanced signal detection method in which the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$2^{\text {nd}}$ </tex-math></inline-formula> harmonics of the carrier frequency ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\omega _{\textit {ref}}$ </tex-math></inline-formula> ) is applied. As opposed to conventional LIA, this method reduces flicker noise and input offset effects that limit the signal-to-noise ratio (SNR). In terms of signal power detection, the proposed single-channel LIA architecture does not require phase synchronization between input and reference signals. The circuit is designed for a reference frequency of 100 Hz, suitable for biomedical applications. The LIA is Implemented using 0.18- <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu \text{m}$ </tex-math></inline-formula> CMOS technology with a single power supply voltage of 1.8 V. The system consumes 360 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu \text{W}$ </tex-math></inline-formula> at an operating frequency of 100 Hz and presents a high input sensitivity dynamic range for a detection bandwidth of 50 Hz and a FOM of 116 dB.

An improved reverse flow injection analysis (rFIA) technique for determination of nanomolar concentrations of ammonium in natural waters with automatic background fluorescence detection: Ammonification during a Karenia brevis bloom in Tampa Bay

Ammonium is the most energetically favorable form of inorganic nitrogen that phytoplankton can take up, and its availability often limits phytoplankton biomass. Recently, the majority of methods for quantification of ammonium at nanomolar concentrations have used the reaction between ammonium and o-phthaldialdehyde (OPA). Two different approaches have been employed. One involves gas diffusion of ammonia across a membrane. The other entails a more direct approach and utilizes sulfite as a reducing agent to form a fluorescent product. A benefit of the separation technique is that natural fluorescence of the samples does not yield a false positive signal, however membrane failure and clogging are challenging analytical problems. The direct reaction between the analyte and OPA, although devoid of problems associated with membranes, requires a correction for background fluorescence. This work presents the development of a reverse flow injection analysis (rFIA) method with automatic background fluorescence correction and a low detection limit, along with its application to measure underway ammonium concentrations in Tampa Bay during a bloom of the toxic dinoflagellate Karenia brevis with concurrent fish kills in June of 2021. During field testing elevated concentrations of ammonium were found to coincide with the largest number of observed floating dead fish and lethal K. brevis concentrations. The method employs a sulfite-formaldehyde reagent mixed with the sample into which an OPA reagent is injected and heated to activate fluorescence. Fluorescence of the sample is measured before the injection of OPA and at the peak of OPA injection, allowing for differentiation of the background fluorescence from the signal proportional to analyte. The detection limit and the limit of quantitation for this technique are 2.3 nM and 7.5 nM, respectively, and the coefficient of variation is 0.6% for replicate deionized water blanks (N = 10). Advantages of this method compared to techniques capable of both nanomolar detection of ammonium and improved accuracy via background fluorescence correction include a greater sample throughput, an increase in reported useful reagent lifetime, and the ability to quantify ammonium in coastal waters.

High speed visual insect swarm tracker (Hi-VISTA) used to identify the effects of confinement on individual insect flight

Individual insects flying in crowded assemblies perform complex aerial maneuvers by sensing and feeding back neighbor measurements to small changes in their wing motions. To understand the individual feedback rules that permit these fast, adaptive behaviors in group flight, both experimental preparations inducing crowded flight and high-speed tracking systems capable of tracking both body motions and more subtle wing motion changes for multiple insects in simultaneous flight are needed. This measurement capability extends tracking beyond the previous focus on individual insects to multiple insects. This paper describes an experimental preparation that induces crowded insect flight in more naturalistic conditions (a laboratory-outdoor transition tunnel) and directly compares the resulting flight performance to traditional flight enclosures. Measurements are made possible via the introduction of a multi-agent high speed insect tracker called Hi-VISTA, which provides a capability to track wing and body motions of multiple insects using high speed cameras (9000–12 500 fps). Processing steps consist of automatic background identification, data association, hull reconstruction, segmentation, and feature measurement. To improve the biological relevance of laboratory experiments and develop a platform for interaction studies, this paper applies the Hi-VISTA measurement system to Apis mellifera foragers habituated to transit flights through the transparent transition environment. Binary statistical analysis (Welch’s t-test, Cohen’s d effect size) of 95 flight trajectories is presented, quantifying the differences between flights in an unobstructed environment and in a confined tunnel volume. The results indicate that body pitch angle, heading rate, flapping frequency, and vertical speed (heave) are each affected by confinement, and other flight variables show minor or statistically insignificant changes. These results form a baseline as swarm tracking and analysis begins to isolate the effects of neighbors from environmental enclosures, and improve the connection of high speed insect laboratory experiments to outdoor field experiments.

Automated ExEm-spFRET Microscope.

Excitation–emission-spectral unmixing-based fluorescence resonance energy transfer (ExEm-spFRET) microscopy exhibits excellent robustness in living cells. We here develop an automatic ExEm-spFRET microscope with 3.04 s of time resolution for a quantitative FRET imaging. The user-friendly interface software has been designed to operate in two modes: administrator and user. Automatic background recognition, subtraction, and cell segmentation were integrated into the software, which enables FRET calibration or measurement in a one-click operation manner. In administrator mode, both correction factors and spectral fingerprints are only calibrated periodically for a stable system. In user mode, quantitative ExEm-spFRET imaging is directly implemented for FRET samples. We implemented quantitative ExEm-spFRET imaging for living cells expressing different tandem constructs (C80Y, C40Y, C10Y, and C4Y, respectively) and obtained consistent results for at least 3 months, demonstrating the stability of our microscope. Next, we investigated Bcl-xL-Bad interaction by using ExEm-spFRET imaging and FRET two-hybrid assay and found that the Bcl-xL-Bad complexes exist mainly in Bad-Bcl-xL trimers in healthy cells and Bad-Bcl-xL2 trimers in apoptotic cells. We also performed time-lapse FRET imaging on our system for living cells expressing Yellow Cameleon 3.6 (YC3.6) to monitor ionomycin-induced rapid extracellular Ca2+ influx with a time interval of 5 s for total 250 s.