Abstract: A method for detecting urine formed elements based on deep learning and context relation includes performing video capture on a labeled urine sample to obtain a labeled urine video; extracting an image from the labeled urine video to obtain single-frame image data with a time sequence n; and labeling position and category of a formed element on the single-frame image data to obtain a formed element dataset, and dividing the formed element dataset into a training set and a testing set. The method also includes constructing a urine formed element detection and classification model, and comparing and identifying functions, and using the training set and the testing set to train the model. The method further includes performing video capture on a to-be-detected urine sample in the visual field of the microscopic imaging device to obtain a to-be-detected urine video.

Abstract: The present invention relates to a method of preventing abnormal large grains from being included in a NCD thin film during a hot filament CVD process by appropriately controlling the deposition condition regarding a temperature-measuring means, a deposition pressure, an electrical potential and/or the composition of a raw material gas flow.

Abstract: The present invention relates to a method of preventing abnormal large grains from being included in a NCD thin film during a hot filament CVD process by appropriately controlling the deposition condition regarding a temperature-measuring means, a deposition pressure, an electrical potential and/or the composition of a raw material gas flow.