scholarly journals Human-in-the-Loop Segmentation of Earth Surface Imagery

2021 ◽  
Author(s):  
Daniel Buscombe ◽  
Evan Goldstein ◽  
Chris Sherwood ◽  
Cameron Bodine ◽  
Jenna Brown ◽  
...  
Keyword(s):  
Supervised Machine Learning ◽  
Earth Surface ◽  
Widespread Application ◽  
Spectral Signatures ◽  
Human In The Loop ◽  
Automated Method ◽  
Mixed Pixels ◽  
Scientific Potential ◽  
Manual Methods ◽  
River Estuarine

Segmentation, or the classification of pixels (grid cells) in imagery, is ubiquitously applied in the natural sciences. Manual methods are often prohibitively time-consuming, especially those images consisting of small objects and/or significant spatial heterogeneity of colors or textures. Labeling complicated regions of transition that in Earth surface imagery are represented by collections of mixed-pixels, -textures, and -spectral signatures, can be especially error-prone because it is difficult to reliably unmix, identify and delineate consistently. However, the success of supervised machine learning (ML) approaches is entirely dependent on good label data. We describe a fast, semi-automated, method for interactive segmentation of N-dimensional (x,y,N) images into two-dimensional (x,y) label images. It uses human-in-the-loop ML to achieve consensus between the labeler and a model in an iterative workflow. The technique is reproducible; the sequence of decisions made by human labeler and ML algorithms can be encoded to file, so the entire process can be played back and new outputs generated with alternative decisions and/or algorithms. We illustrate the scientific potential of segmentation of imagery of diverse settings and image types using six case studies from river, estuarine, and open coast environments. These photographic and non-photographic imagery consist of 1- and 3-bands on regular and irregular grids ranging from centimeters to tens of meters. We demonstrate high levels of agreement in label images generated by several labelers on the same imagery, and make suggestions to achieve consensus and measure uncertainty, ideal for widespread application in training supervised ML for image segmentation.

Spatial Roadway Condition-Assessment Mapping Utilizing Smartphones and Machine Learning Algorithms

2021 ◽  
pp. 036119812110061
Author(s):  
Charalambos Kyriakou ◽  
Symeon E. Christodoulou ◽  
Loukas Dimitriou
Keyword(s):  
Machine Learning ◽  
Learning Algorithms ◽  
Condition Assessment ◽  
Machine Learning Algorithms ◽  
Supervised Machine Learning ◽  
Geographical Information ◽  
Related Field ◽  
Pavement Surface ◽  
Automated Method ◽  
Smartphone Technology

The paper presents a data-driven framework and related field studies on the use of supervised machine learning and smartphone technology for the spatial condition-assessment mapping of roadway pavement surface anomalies. The study explores the use of data, collected by sensors from a smartphone and a vehicle’s onboard diagnostic device while the vehicle is in movement, for the detection of roadway anomalies. The research proposes a low-cost and automated method to obtain up-to-date information on roadway pavement surface anomalies with the use of smartphone technology, artificial neural networks, robust regression analysis, and supervised machine learning algorithms for multiclass problems. The technology for the suggested system is readily available and accurate and can be utilized in pavement monitoring systems and geographical information system applications. Further, the proposed methodology has been field-tested, exhibiting accuracy levels higher than 90%, and it is currently expanded to include larger datasets and a bigger number of common roadway pavement surface defect types. The proposed system is of practical importance since it provides continuous information on roadway pavement surface conditions, which can be valuable for pavement engineers and public safety.

scholarly journals The results of external quality assessment programme on urine leukocyte and erythrocyte counting in Poland

2020 ◽  
Vol 30 (2) ◽  
pp. 278-286
Author(s):  
Agnieszka Ćwiklińska ◽  
Barbara Kortas-Stempak ◽  
Maciej Jankowski ◽  
Gabriela Bednarczuk ◽  
Aleksandra Fijałkowska ◽  
...  
Keyword(s):  
Quality Assessment ◽  
External Quality Assessment ◽  
Particle Analysis ◽  
External Quality ◽  
Automated Method ◽  
Percentage Difference ◽  
Target Values ◽  
Important Diagnostic Tool ◽  
Manual Methods

Introduction: Urine particle analysis is an important diagnostic tool. The aim of this study was to evaluate the quality of urine leukocyte (WBC) and erythrocyte (RBC) counting results obtained with manual and automated methods in Polish laboratories participating in the international external quality assessment (EQA) programme. Materials and methods: 1400 WBC and RBC counting results were obtained from 183 laboratories in EQA surveys organised by Labquality (Helsinki, Finland) from 2017 to 2019. The between-laboratory coefficient of variation (CV), the percentage difference between the laboratories' results and target values (Q-score (%)), as well as modified Youden plots were analysed. Results: For automated method groups, the medians of inter-laboratory CVs varied from 14% to 33% for WBC counting and from 10% to 39% for RBC counting. For manual method groups, the medians of CV varied from 53% to 71% (WBC) and from 55% to 70% (RBC), and they were significantly higher, in comparison to CVs for most automated method groups (P < 0.001). The highest percentage of results outside the target limits (36%) and the highest range of Q-score (%) (from - 93% to 706%) were observed for laboratories which participated in the surveys for the first or second time. The percentage of deviating results and the ranges of Q-score decreased with an increased frequency of laboratories’ participation in the surveys. Conclusions: The quality of manual methods of urine WBC and RBC counting is unsatisfactory. There is an urgent need to take actions to improve laboratories’ performance and to increase harmonisation of the results.

scholarly journals Development of a semi‐automated method for tumour budding assessment in colorectal cancer and comparison with manual methods

2021 ◽  
Author(s):  
Natalie C Fisher ◽  
Maurice B Loughrey ◽  
Helen G Coleman ◽  
Melvin D Gelbard ◽  
Peter Bankhead ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Automated Method ◽  
Manual Methods ◽  
Tumour Budding

scholarly journals A novel automated approach for improving standardization of the marble burying test enables quantification of burying bouts and activity characteristics

2021 ◽  
Author(s):  
Lucas F. Wahl ◽  
A. Mattijs Punt ◽  
Tara Arbab ◽  
Ingo Willuhn ◽  
Ype Elgersma ◽  
...  
Keyword(s):  
Mouse Models ◽  
Learning Algorithm ◽  
Methodological Approach ◽  
Accurate Method ◽  
Supervised Machine Learning ◽  
Mouse Behavior ◽  
Obsessive Compulsive ◽  
Compulsive Disorder ◽  
Automated Method ◽  
Marble Burying

The marble burying test is a commonly used paradigm to screen phenotypes in mouse models of neurodevelopmental and psychiatric disorders. The current methodological approach relies solely on reporting the number of buried marbles at the end of the test. By measuring the proxy of the behavior (buried marbles), rather than the behavior itself (burying bouts), many important characteristics regarding the temporal aspect of this assay are lost. Here we introduce a novel, automated method to quantify mouse behavior throughout the duration of the marble burying test with the focus on the burying bouts. Using open-source software packages, we trained a supervised machine learning algorithm (the classifier) to distinguish burying behavior in freely moving mice. In order to confirm the classifier's accuracy and uncover the behavioral meaning of the marble burying test, we performed marble burying test in three mouse models: Ube3am-/p+ (Angelman Syndrome model), Shank2-/- (autism model), and Sapap3-/- (obsessive-compulsive disorder model) mice. The classifier scored burying behavior accurately and consistent with the literature in the Ube3am-/p+ mice, which showed decreased levels of burying compared to controls. Shank2-/- mice showed a similar pattern of decreased burying behavior, which was not found in Sapap3-/- mice. Tracking mouse behavior throughout the test enabled us to quantify activity characteristics, revealing hypoactivity in Ube3am-/p+ and hyperactivity in the Shank2-/- mice, indicating that mouse activity is unrelated to burying behavior. Together, we demonstrate that our classifier is an accurate method for the analysis of the marble burying test, providing more information than the currently used methods.

scholarly journals Development of a semi-automated method for tumor budding assessment in colorectal cancer and comparison with manual methods

2021 ◽  
Author(s):  
Natalie C Fisher ◽  
Maurice B Loughrey ◽  
Helen G Coleman ◽  
Melvin D Gelbard ◽  
Peter Bankhead ◽  
...  
Keyword(s):  
Clinical Pathology ◽  
Tumor Budding ◽  
Strongly Correlated ◽  
Colon Cancers ◽  
Automated Method ◽  
Discrete Area ◽  
Prognostic Feature ◽  
Manual Methods ◽  
Automate Assessment ◽  
Tumor Edge

Tumor budding is an established prognostic feature in multiple cancers but routine assessment has not yet been incorporated into clinical pathology practice. Recent efforts to standardize and automate assessment have shifted away from haematoxylin and eosin (H&E)-stained images towards cytokeratin (CK) immunohistochemistry. In this study, we compare established manual H&E and cytokeratin budding assessment methods with a new, semi-automated approach built within the QuPath open-source software. We applied our method to tissue cores from the advancing tumor edge in a cohort of stage II/III colon cancers (n=186). The total number of buds detected by each method, over the 186 TMA cores, were as follows; manual H&E (n=503), manual CK (n=2290) and semi-automated (n=5138). More than four times the number of buds were detected using CK compared to H&E. A total of 1734 individual buds were identified both using manual assessment and semi-automated detection on CK images, representing 75.7% of the total buds identified manually (n=2290) and 33.7% of the total buds detected using our proposed semi-automated method (n=5138). Higher bud scores by the semi-automated method were due to any discrete area of CK immunopositivity within an accepted area range being identified as a bud, regardless of shape or crispness of definition, and to inclusion of tumor cell clusters within glandular lumina ('luminal pseudobuds'). Although absolute numbers differed, semi-automated and manual bud counts were strongly correlated across cores (ρ=0.81, p<0.0001). Despite the random, rather than 'hotspot', nature of tumor core sampling, all methods of budding assessment demonstrated poorer survival associated with higher budding scores. In conclusion, we present a new QuPath-based approach to tumor budding assessment, which compares favorably to current established methods and offers a freely-available, rapid and transparent tool that is also applicable to whole slide images.

Total and direct-reacting bilirubin values by automated methods compared with liquid chromatography and with manual methods for determining delta bilirubin

1990 ◽  
Vol 36 (5) ◽  
pp. 788-791 ◽  
Author(s):  
P Rosenthal ◽  
M T Keefe ◽  
D Henton ◽  
M Cheng ◽  
C R Lee ◽  
...  
Keyword(s):  
Liquid Chromatography ◽  
Bile Acids ◽  
Total Bilirubin ◽  
Clinical Use ◽  
Serum Samples ◽  
Manual Method ◽  
Automated Method ◽  
Manual Methods ◽  
Automated Methods ◽  
Conjugated Bilirubin

Abstract This study compares total and direct-reacting bilirubin values in 40 serum samples from patients with various diagnoses, as measured by automated methods (Beckman Synchron CX-5, Beckman Astra 8, Kodak Ektachem 700) and HPLC and by a manual method for delta bilirubin. For total bilirubin, within-run CVs were less than 6%. The Ektachem 700 method underestimated bilirubin with serum samples from patients with Crigler-Najjar syndrome and from newborns in whom unconjugated bilirubin concentrations were increased but conjugated bilirubins were not present or were present only in small amounts. The Astra 8 and Synchron CX-5 methods were inaccurate with cholestatic serum samples, in which conjugated bilirubin concentrations were increased and other compounds such as bile acids could be expected to interfere. We conclude that each automated method examined provides reasonable estimates for total and direct-reacting bilirubin values for routine clinical use. The need for each laboratory to select the appropriate bilirubin method for its particular situation is obvious.

scholarly journals Comparative Analysis of Erythrocyte Sedimentation Rate Measured by Automated and Manual Methods in Anaemic Patients

2020 ◽  
Vol 12 (04) ◽  
pp. 239-243
Author(s):  
Vikram Narang ◽  
Sumit Grover ◽  
Amandeep Kaur Kang ◽  
Avantika Garg ◽  
Neena Sood
Keyword(s):  
Erythrocyte Sedimentation Rate ◽  
Sedimentation Rate ◽  
Manual Measurement ◽  
Automated Method ◽  
Erythrocyte Sedimentation ◽  
Manual Methods ◽  
Automated Instrument ◽  
Very High ◽  
Edta Blood ◽  
High Range

Abstract Purpose Erythrocyte sedimentation rate (ESR) is a widely used indicator of inflammation and a routinely done hematology investigation to monitor patients of autoimmune and infectious diseases. We aimed to compare the ESR results obtained by Roller 20LC automated instrument and standard reference Westergren method and analyzed the effect of anemia (hematocrit) on ESR measurements through the automated method. Methods We analyzed 1377 random anemic OPD patients (hematocrit [HCT] < 35%) for ESR levels measured by Roller 20LC using EDTA blood and Westergren method using citrated blood for a one and half year period from January 1, 2018 to June 30, 2019. Fabry’s formula was used to correct the Westergren ESR. Results The total number of samples after evaluation were divided into low (n = 232), intermediate (n = 417), high (n = 406), and very high range of ESR (≥100 mm/hr; n = 422). Mean difference between values of corrected and automated ESR for the low, intermediate, high and very high ESR range was 2.33 ± 5.03, 10.95 ± 8.04, 28.22 ± 19.11 and 43.3 ± 19.22 mm/hr, respectively. The 95% limit of agreement calculated by the Bland–Altmann analysis between the two methods for low-ESR range was −7.53 to 12.2 (highest correlation coefficient –0.65), while for very high ESR, range was −5.1 to 81.5 (least coefficient of 0.18) (p < 0.001). Conclusion In laboratories with high-sample load and where manual measurement may be tedious, the automated method of ESR measurement can safely replace the Westergren method for low-ESR values in patients with low hematocrit. While for high-ESR values, validation by the standard Westergren method may be needed.

"True" Creatinine Chromogen Determination in Serum and Urine by Semi-automated Analysis

1965 ◽  
Vol 11 (8) ◽  
pp. 763-770 ◽  
Author(s):  
E Polar ◽  
J Metcoff
Keyword(s):  
Cation Exchange ◽  
Phosphate Buffer ◽  
Exchange Resin ◽  
Cation Exchange Resin ◽  
Automated Analysis ◽  
Positive Error ◽  
Strong Cation Exchange ◽  
Automated Method ◽  
Manual Methods ◽  
Serum And Urine

Abstract The Jaffè reaction, used to determine creatinine in urine and plasma or serum, gives a positive error in the latter because of the interference of noncreatinine. In man (particularly children) endogenous creatinine normally occurs at a relatively low concentration in serum (0.3-1.0 mg./100 ml.). The analytic error contributed by noncreatinine chromogens, therefore, may be considerable. Of various manual methods to eliminate this none has been used with automated creatinine analyses, which generally measure total creatinine-like chromogen. In the present investigation a strong cation-exchange resin is used manually for rapidly adsorbing and desorbing "true" creatinine from plasma which is then analyzed by an automated method. A phosphate buffer (pH 12.4) is substituted for 5% (w/v) NaOH, dialysis eliminated, and the instrument manifold design is modified. Concentrations of "true" creatinine from 0.2-2.0 mg./100 ml. can be determined at a rate of 20 samples per hr. in serum or 40/hr. in urine, with a recovery of 98-101%.

scholarly journals 1786. An Automated Method to Assess Oligonucleotide Primer and Probe Complementarity to Genomic Targets in Infectious Disease qPCR Assays

2019 ◽  
Vol 6 (Supplement_2) ◽  
pp. S657-S658
Author(s):  
Rohita Sinha ◽  
Mark Wissel ◽  
Katelyn Bartlett ◽  
James Grantham ◽  
Steve Kleiboeker
Keyword(s):  
Traditional Approach ◽  
Computing Time ◽  
Genomic Sequences ◽  
Oligonucleotide Primer ◽  
Automated Method ◽  
Pathogen Variation ◽  
Sequence Complementarity ◽  
Taqman Pcr ◽  
Manual Methods ◽  
High Degree

Abstract Background Success of real-time TaqMan PCR (qPCR) in detecting pathogen targets and quantifying pathogen load is dependent upon frequent assay monitoring. This is due to i) the high degree of complementarity needed between primers / probes and genomic targets for assay accuracy and ii) natural pathogen variation and evolution. Failure to monitor and refine may result in false negativity or under quantification. Here we present a bioinformatics tool to identify potential problems resulting from newly discovered genomic mutations in primer/probe regions. Methods The tool performs an unbiased and automated search of the NCBI database, collects relevant genomic sequences based on user-defined Taxon-ID and executes a Python program to discard synthetic sequences. A profile of primer-probe sequence complementarity to targets is then generated. While the tool can be used for any microbe, here we present results for our laboratory’s cytomegalovirus (CMV) qPCR primer-probe analysis. In addition, our laboratory’s traditional approach utilizing alignment software was performed (download of all CMV sequences (~10,000) followed by iterative alignment building of these against our primers and probes). The amount of time to perform the automated and manual methods was recorded. Results The tool retrieved 8,732 sequences from NCBI and compared these to the CMV qPCR primers and probes. The tool found 2,501 alignments between the primers / probes and the downloaded genomic data (~15 minutes to finish (6 CPUs)). A total of 64% (1,624/2,501) of BLASTn alignments were exact matches between all primers / probes and viral genomic sequences. 17.5% (439/2,501) of alignments had 1 mismatch at either 5’ or 3’ terminus, and 1% (25/2,501) of alignments had two mismatches with the primers / probes. Similar results were found using a primarily manual approach (which took approx. 5 hours computing time and 20 hours of labor). Conclusion This new bioinformatics approach performed indistinguishably vs. a manual approach and did so in minutes rather than days. Both methods led to the conclusion that, by virtue of our design involving overlapping primers and probes, none of the identified mismatches are predicted to lead to false negativity or under quantification in our current CMV qPCR assay. Disclosures All authors: No reported disclosures.

scholarly journals COMPARISON BETWEEN FOUR USUAL METHODS OF IDENTIFICATION OF Candida SPECIES

2015 ◽  
Vol 57 (4) ◽  
pp. 281-287 ◽  
Author(s):  
Margarida Neves SOUZA ◽  
Stéfanie Otowicz ORTIZ ◽  
Marcelo Martins MELLO ◽  
Flávio de Mattos OLIVEIRA ◽  
Luiz Carlos SEVERO ◽  
...  
Keyword(s):  
Candida Species ◽  
Germ Tube ◽  
Mortality Rates ◽  
Correct Identification ◽  
Kappa Index ◽  
Candida Spp ◽  
Chromogenic Medium ◽  
Tube Test ◽  
Automated Method ◽  
Manual Methods

SUMMARY Infection by Candidaspp. is associated with high mortality rates, especially when treatment is not appropriate and/or not immediate. Therefore, it is necessary to correctly identify the genus and species of Candida. The aim of this study was to compare the identification of 89 samples of Candida spp. by the manual methods germ tube test, auxanogram and chromogenic medium in relation to the ID 32C automated method. The concordances between the methods in ascending order, measured by the Kappa index were: ID 32C with CHROMagar Candida(κ = 0.38), ID 32C with auxanogram (κ = 0.59) and ID 32C with germ tube (κ = 0.9). One of the species identified in this study was C. tropicalis,which demonstrated a sensitivity of 46.2%, a specificity of 95.2%, PPV of 80%, NPV of 81.1%, and an accuracy of 80.9% in tests performed with CHROMagar Candida;and a sensitivity of 76.9%, a specificity of 96.8%, PPV of 90.9%, NPV of 91%, and an accuracy of 91% in the auxanogram tests. Therefore, it is necessary to know the advantages and limitations of methods to choose the best combination between them for a fast and correct identification of Candidaspecies.

Sign in / Sign up

Export Citation Format

Share Document