The presence of porosity within magnesium-based orthopaedic implants is known to be beneficial, promoting cell proliferation and vascularisation. However, the presence of porosity increases the surface area available for corrosion, compounding the issue of high corrosion rates which has long been plaguing magnesium-based materials. This work looks at the influence of hydroxyapatite and phosphate conversion coatings on the corrosion performance of conventionally cast, dense Mg-Zn-Zr alloys and binder jet additive manufactured porous Mg-Zn-Zr scaffolds. The performance of coating on dense Mg-Zn-Zr was found to be more effective than the coating on the porous Mg-Zn-Zr scaffold, with the discrepancies attributed to both the microstructure and geometric influence of the binder jet additive manufactured, porous Mg-Zn-Zr scaffold, which not only increases the rate of hydrogen evolution but also reduces the ability of the hydrogen gas generated within the pore channels to escape to the sample’s surface. This restricts the effectiveness of coating application for porous Mg scaffold. Furthermore, the limited diffusion within the pore channels can also result in differing localized corrosion environments, causing discrepancies between the localised corrosion environment within the pore channels and that at the bulk electrolyte.

High-resolution remote sensing (HRRS) images contain complex visual contents and rich detailed information. This paper proposes a multi-level feature fusion method to improve the feature representation for HRRS image retrieval. Firstly, in order to obtain the multi-scale information of HRRS images, mid-level features and high-level features of VGG16 and GoogLeNet are extracted with different input sizes. Then a feature transformation method is proposed to adjust the size and the number of different feature maps, so that the distinct mid-level features and distinct high-level features can be fused separately using element-wise addition. There is a large amount of redundancy in the fusion features, thus small-region max-pooling method is adopted to aggregate the mid-level fusion feature, and global max-pooling method is used to aggregate the high-level fusion feature. Finally, an adaptive weight allocation method based on similarity is proposed to combine mid-level feature and high-level feature. Experimental results show that the multi-level feature fusion is an effective method to enhance the feature representation, thereby improving the retrieval performance.

SPIE Digital Library Proceedings