Search

Purpose: To develop an automated method to delineate lens epithelial cells and to quantify expression of fluorescent signal of biomarkers in each nucleus and cytoplasm of lens epithelial cells in a histological section. Methods: An automated algorithm was developed in Matlab™ to localize and quantify fluorescence signal in lens epithelial cells in histological images. A region of interest representing the lens epithelium was manually demarcated in each input image. Individual cell nuclei within the region of interest were automatically delineated based on watershed segmentation and thresholding. Fluorescence signal was quantified within nuclei and cytoplasms. The classification of fluorescence signal was based on local background. Classification of cells as labelled or not labelled was thereafter optimized as compared to visual classification of a limited dataset. The performance of the automated classification was evaluated by asking eleven independent blinded observers to classify all cells (n=395) in one lens image. Time consumed by the automatic algorithm and visual /manual classification of nuclei, was recorded. Results: On an average, 77 % of the cells were correctly classified as compared to the majority vote of the visual observers. The average agreement among visual observers was 83 %. However, variation among visual observers was high, and agreement between two visual observers was as low as 71 % in the worst case. Automated classification was on average 10 times faster than manual scoring. Conclusion: The presented method enables objective and fast detection of lens epithelial cells and quantification of expression of fluorescent signal in a histological section of rat lens, with accuracy comparable to the variability between different visual observers. Furthermore, automated scoring is unbiased and reproducible, and results in a 10-fold increase in throughput.

The assessment of the risk maps for the seismic vulnerability at large scale is based on the vulnerability of each building. In order to determine these vulnerabilities, it is first required to assign to each building its construction class. The construction class is needed to define the seismic behavior of the building. Since the structures have been built in different times, this construction classes are different between them, and their vulnerability values have been pre-defined in order to simplify the analysis. If, for a small city the task of assign to each building its construction class can be done by hand, in a big city, with thousands of buildings, this process can result long and hard. In this Master project the possibility of assigning the construction classes by using a machine learning method will be studied. This method use as a data set of the city of Lausanne from previous works which contains around 1000 buildings analyzed by hand with their construction classes. The goal of this method is to find the relationships between the attributes of the buildings to their construction classes. Thanks to the statistical offices, it is possible to obtain the attributes in an automatic way for every building in a city. Based on the relations that the machine learning method finds (and using the statistical attributes) the construction class for every building can finally be determined, in a quick and efficient manner. Once obtained the vulnerability’s values for each building, the risk maps can be drawn. In this project there will be studied three maps, two with the Europeans’ typologies (LM1, LM2), and one with the Swiss’ typologies (UniGE). Some comparisons will be made between the maps, in order to highlight the differences. Some other comparisons will be made to show the impact of the new micro zonation of Lausanne and of the new optimized method for the determination of the seismic displacement demand (N2 optimized method) on the maps. This is done for the mechanical methods (LM2 and UniGE) only. Finally, there will be shown the relationships between the soil’s characteristics (micro zonation), the building’s attributes and the city’s historical evolution with the risk maps.

Highlights • Inaccuracies of QRS and T-wave markers significantly impact QRS-Ta estimation. • These errors influence the classification of clinically relevant abnormal values. • Our algorithm provides robust measurements in the presence of inaccurate VCG markers. • We present for the first time, the distribution of the QRS-Ta in a large cohort., The spatial QRS-T angle (QRS-Ta) derived from the vectorcardiogram (VCG) is a strong risk predictor for ventricular arrhythmia and sudden cardiac death with potential use for mass screening. Accurate QRS-Ta estimation in the presence of ECG delineation errors is crucial for its deployment as a prognostic test. Our study assessed the effect of inaccurate QRS and T-wave marker placement on QRS-Ta estimation and proposes a robust method for its calculation. Reference QRS-Ta measurements were derived from 1,512 VCGs manually annotated by three expert reviewers. We systematically changed onset and offset timings of QRS and T-wave markers to simulate inaccurate placement. The QRS-Ta was recalculated using a standard approach and our proposed algorithm, which limits the impact of VCG marker inaccuracies by defining the vector origin as an interval preceding QRS-onset and redefines the beginning and end of QRS and T-wave loops. Using the standard approach, mean absolute errors (MAE) in peak QRS-Ta were >40% and sensitivity and precision in the detection of abnormality (>105°) were 15 ms. Using our proposed algorithm, MAE for peak QRS-Ta were reduced to 94% for inaccuracies up to ±15 ms. Similar results were obtained for mean QRS-Ta. In conclusion, inaccuracies of QRS and T-wave markers can significantly influence the QRS-Ta. Our proposed algorithm provides robust QRS-Ta measurements in the presence of inaccurate VCG annotation, enabling its use in large datasets.