Microstructural Analysis of Ceramic Materials Through Machine Learning Techniques: A Benchmarking Interdisciplinary Review for Ceramic Material Characterization

This study presents a revolutionary analytical study on ceramic microstructural images in the context of sophisticated image processing as well as machine learning techniques for automated and most accurate characterization of ceramic material properties. The main objective here is to scientifically correlate the derived microstructural image features with the concern material properties and this can be purely governed by specific material compositions that lead to better productivity. So, this particular unique study purely focuses on some efficient machine learning algorithms like K-means and fuzzy C-means in contrast to the traditional image processing concepts like Canny and Sobel for most effective classification as well as automated evaluation of ceramic microstructural image features. In addition, more advanced deep learning techniques like convolutional neural network also reflected here with in depth comparative analysis in view of its superiority for crack estimation, pore characterization, phase identification, grain boundary detection and grain area calculation that leads to accurate material characterizations. Further, continuous up gradations are still highlighted here with recent advanced deep learning models like Auto encoders, generative adversarial networks and recurrent neural networks for much more accurate characterization of ceramic microstructural image features boosting both material design and productivity in great manner. Apart from that, some latest image property related transformations are also investigated here for accurately tracing the critical image features while this study even adopts some exhaustive bibliometric analysis to precisely identify the effective microstructural evaluation methods from the existing literature. Hence this novel study not only provides a concrete pathway for identifying the most effective mechanism for ceramic microstructural image property evaluation and characterization in the present scenario but it also emphasizes the potential future research roadmap in this domain.