Search

With the development of deep learning, breakthroughs have been made in the field of semantic segmentation. However, it is difficult to generate a fine mask on the same medical images because medical images have low contrast, high resolution, and insufficient semantic information. In most scenarios, existing approaches mostly use a pooling layer to reduce the resolution of feature maps. Therefore, it is difficult for them to consider the whole image features, resulting in information loss and performance degradation. In this paper, a multiscale asymmetric encoder-decoder semantic segmentation network is proposed. The network consists of two parts, which perform feature extraction and image restoration on the input, respectively. The encoder network obtains multiscale feature information by connecting multiple ASPP modules to form a feature pyramid. Meanwhile, the upsampling layer of each decoder can be connected to the feature map generated by the corresponding ASPP module. Finally, the classification information of each pixel is obtained through the sigmoid function. The performance of the proposed method can be verified on publicly available datasets. The experimental evidence shows that the proposed method can take full advantage of multiscale feature information and achieve superior performance with less inference computational cost.

In this paper, a Micro-Hole Array Sprayer-Assisted LIBS (MHAS-LIBS) method for liquid sample analysis was discussed. The experimental setup of this method was introduced and the stability, sensitivity and quantitative analysis properties of LIBS signal were studied. The experimental results showed that the RSD value of the spectral signal of metal elements in liquid samples was less than 2.2%. The LODs of Na, Ca, Mg and K were 0.67 mg/L, 0.29 mg/L, 0.85 mg/L, 6.18 mg/L respectively. Compared with other LIBS signal enhancement methods, the LODs of different elements were comparable or even better. The application of quantitative analysis method of MHAS-LIBS in the range of possible concentrations in sea water samples was discussed. Especially for the elements showing nonlinear characteristics, three concentration fitting methods, including the Scheibe-Lomakin function and the Cubic function, were compared. The results showed that it was a simple and practical method to fit the experimental data with the Cubic function in the target application. Finally, a quantitative method for seawater salinity determination was proposed and the estimated salinities of 5 different samples were approximately the same as the actual salinity, with a minimum error of 1.28%.

In order to develop a method to analyze metal elements in thin-film samples rapidly, directly and without sample preparation, and to understand the mechanism of laser-film interaction and plasma formation and evolution, a laboratory laser-induced breakdown spectroscopy system was established recently for nanometer-film analysis. ZrO(2) films prepared on silicon chips by a sol-gel process were employed in the following experiment and their thickness was about 40 nm. By the initial investigation that we carried out, the stability of this system was verified and the relative standard deviation of the target peak was found to be lower than 1.6% with the help of a position system. The influences of different experimental parameters, such as laser energy, laser focus to sample distance (LFTSD) settings, and gate delay, were studied under conditions of room temperature and atmospheric pressure. The experimental results show that the LFTSD was one of the most important parameters for plasma formation and spectral collection in comparison with other parameters by means of plasma spectra and images. So the effects of the LFTSD on the spectra, plasma evolution, and craters are specially discussed in this paper. At last, we calculated the plasma temperature and electron density under optimal parameters for quantitative analysis. The result shows that the established system is available for qualitative and quantitative analysis of films under conditions of single pulse and low ablation energy.