3 resultados para Human decisions
em Repositório Institucional UNESP - Universidade Estadual Paulista "Julio de Mesquita Filho"
Resumo:
An intelligent system that emulates human decision behaviour based on visual data acquisition is proposed. The approach is useful in applications where images are used to supply information to specialists who will choose suitable actions. An artificial neural classifier aids a fuzzy decision support system to deal with uncertainty and imprecision present in available information. Advantages of both techniques are exploited complementarily. As an example, this method was applied in automatic focus checking and adjustment in video monitor manufacturing. Copyright © 2005 IFAC.
Resumo:
Considering that counting the percentage of CD4 T lymphocytes can add prognostic information regarding patients infected with HIV, the aim of this study was to evaluate the percentage values of CD4+ T lymphocytes from 81 patients determined by flow cytometry and estimated by flow cytometry in conjunction with a hematology counter. Means were compared through the Student's t-test. Pearson's correlation was determined, and the agreement between results was tested by Bland-Altman. The level of significance was P < 0.05. It was found a significantly higher mean difference between the relative values of CD4+ T lymphocytes to the hematologic counter (P < 0.05), for all strata studied. Positive and significant correlations (P < 0.01) were found between the strata CD4 < 200 cells/mL (r = 0.93), between 200 and 500 cells/mL (r = 0.65), and >500 cells/mL (r = 0.81). The limits of agreement were 1.0 +/- 3.8% for the stratum of CD4 < 200 cells/mL, approximately 2.2 +/- 13.5% for the stratum of CD4 between 200 and 500 cells/mL, and approximately 6.2 +/- 20.4% for the stratum > 500 cells/mL. The differences in the percentages of CD4+ T lymphocytes obtained by different methodologies could lead to conflict when used in clinical decisions related to the treatment and care of people infected with HIV.
Resumo:
Purpose: To compare visual inspection (VI), radiographic examination (RX) and the laser fluorescence device DIAGNOdent (L), as well as their combinations in vitro regarding treatment decisions for occlusal surfaces. Methods: 72 extracted human permanent teeth (molars and premolars) were used. Treatment decisions were recorded by three calibrated examiners, and the options available were fissure sealant and conservative restoration. For validation of treatment decisions, the teeth were sectioned and examined in a stereomicroscope. Thereafter, dental slices were scanned and the images were edited to facilitate classification of existing carious lesions. Intra and inter-examiner reproducibility for the determination of treatment plans were calculated using Cohen's kappa test (95%-CI). Sensitivity, specificity, positive and negative predictive values, and the area under the ROC curve were also calculated. Results: VI and L provided on average the greatest intra- and inter-examiner reproducibility, respectively. Although the combination of diagnostic methods may decrease both intra- and inter examiners reproducibility, combination of VI, L and RX resulted in the greatest sensitivity, being statistically superior to RX and L. There was more inter-examiner agreement for the option of restorative treatment, while the use of sealants as a treatment option yielded the lowest values. Negative predictive values were numerically inferior to positive predictive values, indicating that the examiners preferred not to restore a carious tooth than to proceed operatively in an intact tooth. The combination of the three methods studied showed the best results in determining treatment plans for occlusal surfaces, when compared to the other types of exams. on the other hand, radiographic examination and laser fluorescence were less efficient when used alone.