Mohammed Yaqoob, Mohammed Ishaq, Mohammed Yusuf Ansari, Issac Sujay Anand John Jayachandran, Mohammed Hashim, Thomas Daniel Seers

Morphological characterization of microcrystalline rock textures typically relies upon the visual interpretation and manual measurement of scanning electron microscopy (SEM) imagery: a practice fraught with subjectivity, inefficiency, sampling bias, and data loss. We introduce a state-of-the-art computer vision pipeline, built on deep learning architectures, for segmenting and classifying individual microcrystals from SEM images. Initially applied to low-Mg calcite carbonate rocks, instance segmentation is achieved using a custom-tuned version of Meta's Segment Anything Model (SAM). To train and test the classifier, we utilized 48 SEM images of diverse carbonate microtextures composed of Low-Mg calcite from studies performed worldwide. Each individual microcrystal (1852 in total) was labelled according to a bipartite classification scheme, encompassing both crystal shape (rhombic, polyhedral, amorphous, and spherical), and degree of crystal facet definition (euhedral to subhedral, anhedral), with a total of four distinct classes. MicroCrystalNet: our proposed classification model employs a convolutional neural network architecture, incorporating advanced feature map processing (feature normalization, dimensionality reduction, and sparse feature selection), integrated within a novel Normalized Sparse Reduction block. Performance metrics reveals excellent Average Precision scores (AP = 0.93-0.98) and Area Under Receiver-Operator Curve values (AUC = 0.95-0.99) across all classes, with visual comparison to manual ground truth images demonstrating powerful inter-class discriminatory power, even in the presence of occlusions.

This study establishes a baseline for the automated classification of microcrystalline rock textures. Leveraging SEM imagery and our high-throughput segmentation and classification framework, we enable quantitative characterization of microcrystalline geologic media. For instance, MicroCrystalNet can analyze microporous carbonate rocks at scale, revealing spatiotemporal trends in microporosity and diagenesis. To support reproducibility and further research, we provide the labeled dataset, feature extraction tool, and deep learning-based pipeline as open-source resources. This framework can be extended to other lithologies or non-geologic microcrystalline materials with the addition of specific training images and labels.