Enhanced Material Classification via MobileSEMNet: Leveraging MobileNetV2 for SEM Image Analysis
Bibliographic record
Abstract
Background: Scanning electron microscopy (SEM) has been instrumental in elucidating material details, enabling the use of SEM images for machine learning applications.This study introduces a novel automated classification model employing deep learning techniques to classify materials using SEM imagery with high accuracy.Materials and Methods: A publicly accessible dataset comprising over 20,000 SEM images across 10 classes was utilized.The dataset was bifurcated into training and testing subsets for model development.The advanced MobileSEMNet model was trained using a pre-trained MobileNetV2 architecture, wherein deep features were extracted via the global average pooling layer.Feature extraction proceeded with fixed-size patch division using the MobileNetV2 network.During feature selection, the neighborhood component analysis (NCA) was employed to distill the most informative 512 features from the feature vector.Subsequently, a classification toolkit determined the Support Vector Machine (SVM) as the optimal shallow classifier.The model was refined using 10-fold cross-validation.Results: The primary aim to surpass the classification accuracy of the standalone MobileNetV2 was achieved; the MobileNetV2 secured a 94.85% accuracy rate, whereas the proposed MobileSEMNet model attained a 96.87% accuracy rate on the test data.Conclusions: MobileSEMNet has proven to be highly effective in the classification of SEM images within a significant dataset, signaling its potential utility in material science.The results unequivocally underscore the model's capacity for automated material type detection, thereby enhancing analytical precision and automation in the field.This work underscores the advantages of deep learning methodologies in advancing material classification in various scientific domains.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".