Deep Learning-Driven Microstructure Characterization and Vickers Hardness Prediction of Mg-Gd Alloys

TL;DR

A multimodal fusion learning framework based on image processing and deep learning techniques that integrates both elemental composition and microstructural features to accurately predict the Vickers hardness of solid-solution Mg-Gd alloys is proposed.

Abstract

In the field of materials science, exploring the relationship between composition, microstructure, and properties has long been a critical research focus. The mechanical performance of solid-solution Mg-Gd alloys is significantly influenced by Gd content, dendritic structures, and the presence of secondary phases. To better analyze and predict the impact of these factors, this study proposes a multimodal fusion learning framework based on image processing and deep learning techniques. This framework integrates both elemental composition and microstructural features to accurately predict the Vickers hardness of solid-solution Mg-Gd alloys. Initially, deep learning methods were employed to extract microstructural information from a variety of solid-solution Mg-Gd alloy images obtained from literature and experiments. This provided precise grain size and secondary phase microstructural features for performance prediction tasks. Subsequently, these quantitative analysis results were combined with Gd content information to construct a performance prediction dataset. Finally, a regression model based on the Transformer architecture was used to predict the Vickers hardness of Mg-Gd alloys. The experimental results indicate that the Transformer model performs best in terms of prediction accuracy, achieving an R^2 value of 0.9. Additionally, SHAP analysis identified critical values for four key features affecting the Vickers hardness of Mg-Gd alloys, providing valuable guidance for alloy design. These findings not only enhance the understanding of alloy performance but also offer theoretical support for future material design and optimization.

Figures (13)

• Figure 1: Overall framework of this study (revealing the relationship between material images, material microstructure, and material properties).
• Figure 2: Edge detection model structure. The entire model framework consists of three stages. (a) presents the general structure of the edge extraction module; (b) is multi-scale adaptive feature refinement module; (c) is based on the compact part convolution module; (d) is based on channel and spatial large-kernel attention module.
• Figure 3: Three instances of pixel difference convolution derived from extended LBP descriptors. One can derive other instances by designing the picking strategy of the pixel pairs.
• Figure 4: The edge detection model results at each stage are as follows: (a) represents the original image; (b) displays the grain boundary results detected in the first stage; (c) shows the grain boundary results detected in the second stage; (d) illustrates the grain boundary results detected in the third stage, and (e) presents the grain boundary results fused from the features of all three stages.
• Figure 5: Detailed process of the edge repair module. The module utilizes image gradient information and edge detection structure to construct a mask, obtaining the final result using morphological methods from computer vision.