A DenseNet-Based Deep Learning Approach for Accurate Classification and Interpretability in the Diagnosis of Acute Lymphoblastic Leukemia
Bibliographic record
Abstract
Acute Lymphoblastic Leukemia (ALL) is a life- threatening hematologic malignancy and requires precise and early diagnosis. In this work we introduce a deep learning model, based on DenseNets, to fully automate the classification and interpretation of all microscopic images of blood smears of ALL. The densely connected nature of DenseNet boosts gradient flow and feature reuse, providing the solid framework of images-based medical diagnosis. The data that is used in the given study is a balanced dataset due to a fair amount of normal and abnormal white blood cell (WBC) images as can be observed through the class distribution plot. Exploratory data analysis consisting of pixel intensity histograms, dimensionality reduction through PCA and t-SNE showed that the distributions of normal and the leukemic cells can be distinguished by their clustering patterns before the model training. It is seen that there exists a discriminative power in the extracted features. The suggested model had a quick convergence with the training and validation sets of 20 epochs. The training and validation accuracies were stabilized at the values of about 98.5 percent and 98, respectively. Loss values in correspondence also showed a steep decrease, which showed that over fitting had a least extent. The confusion matrix ensures the best classification results of test set with accuracy, precision, recall, and F1-score of 100%, and zero false positive of negative. Also, PCA and t-SNE visualizations prove the ability of the model to draw clear boundaries between different classes. The transparency of predictions can also be seen through the distributions of feature spaces and this allows clinical transparency. Conclusively, the model constructed using DenseNet performed at an impressive level in monitoring ALL through the blood smear pictures, and serves as an encouraging method in assisting the hematologists in their early diagnosis. Future directions can include (1) integration with explanations methods of AI, and (2) testing on multi- center data to check the robustness and generalizability of the method.
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 | 0.001 |
| 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 | 0.001 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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".