https://www.selleckchem.com/products/dynasore.html Then, the classification network is trained by sharing shallow layers of the segmentation network. Testing the proposed method on real patient data shows that it is able to segment the lesion areas in thyroid ultrasound images with a Dice score of 92.65% and to distinguish TFA and FTC with a classification accuracy of 96.00%.Hollow structure and pore size are considered to be crucial to the performance of nitrogen-doped carbon materials. In this paper, a lipstick-like hollow and mesoporous nitrogen-doped carbon (HNC-1000) material is prepared using a bottom-up template participation strategy. The images by scanning electron microscopy and transmission electron microscopy show that the precursor ZnO particles, the intermediate ZnO@ZIF-8 core-shell particles, and the target HNC-1000 particles all maintain a lipstick-like morphology, and HNC-1000 is a hollow nitrogen-doped carbon material. The specific surface area and pore size analyses show that the synthesized HNC-1000 has a very rich mesoporous structure with Vmeso+macro/Vtotal of 94.8% and mean mesopore size at 13.67 nm. X-ray photoelectron spectroscopy results show that the nitrogen in the catalyst HNC-1000 is mainly pyridine nitrogen and graphite nitrogen. The prepared HNC-1000 has excellent ORR catalytic activity with onset potential (0.98 V versus RHE), half-wave potential (0.85 V versus RHE), and limiting current density (5.51 mA cm-2), which is comparable to that of commercial Pt/C (20 wt%) and superior to NC-1000 derived from pristine ZIF-8. HNC-1000 also has good stability and strong methanol tolerance, which is superior to commercial Pt/C catalyst. The improved performance of HNC-1000 is attributed to its hollow and mesoporous morphology. These findings demonstrate a stratage for the rational design and synthesis of practical electrocatalysts.True one-dimensional (1D) van der Waals materials can form two-dimensional (2D) dangling-bond-free anisotropic sur