Transfer Learning-Driven MRI-Based Classification Pipelines for Brain Tumor Diagnosis: Glioma, Meningioma, and Pituitary Tumor Discrimination
Bibliographic record
Abstract
MRI-based classification of brain tumors is an important step in delivering timely and effective treatment. In this study, we tested a method that uses CNNs trained via transfer learning to classify the three main types of brain tumors (gliomas, meningiomas, and pituitary tumors). Mendeley dataset is taken into consideration, containing 6,056 MRI images (2004 brain glioma, 2004 brain meningioma, 2048 brain pituitary). Two types of CNN architectures were tested, AlexNet (trained from scratch) and InceptionV3 (using the weights from ImageNet). All of the images were preprocessed before being fed to the models using image resizing, normalization, and extensive augmentation to ensure accuracy and minimize class imbalance between the tumor categories. Effectively stratified train-test splits of the data allowed for fair performance evaluation of both models. The AlexNet model consistently achieved 94% accuracy, with a precision, recall, and F1-score of 94%, indicating that it could provide reliable performance when classifying brain tumors based on MRI. In contrast, the InceptionV3 model using transfer learning and fine-tuning performed even better than AlexNet, achieving 98% accuracy with a precision, recall, and F1- score of 98%. These results indicate that pre-trained convolutional neural network architectures provide increased classification reliability, significantly reduce training time, and are applicable to medical datasets that contain limited numbers of instances. The findings of this research study illustrate the potential for developing highly accurate and efficient automated deep learning technology to accurately diagnose neuro-oncology diseases using transfer learning. This type of technology will provide a strong basis for Clinical Decision Support Systems (CDSS) that aid radiologists with the interpretation of diagnostic medical images.
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 machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.003 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.002 |
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 source (direct Gemma or distilled Codex), 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".