Whole slide cervical image classification based on convolutional neural network and random forest
Bibliographic record
Abstract
Abstract Cervical cancer is a kind of common female malignancy ranking fourth for mortality worldwide. Traditional histopathological examination, an important diagnosis method of cervical cancer, is still manually performed by pathologists under the microscope, which is labor intensive and error‐prone. In this article, deep learning is used for whole slide cervical image (WSCI) analysis to explore an automatic and effective method for the diagnosis of cervical cancer. We combine convolutional neural network (CNN) and random forest (RF) classifier for whole slide cervical image classification. A new multilevel feature fusion strategy namedEnsembleis used for features extraction from the features extracted by CNN.Ensemblethat fuses features extracted by convolution layers with different depths in CNN together is capable of capturing fusional features which can describe patches at different levels from different convolutional layers. Principal component analysis (PCA) algorithm is introduced for feature reduction of the multilevel features. Our experiments are carried out on WSCI dataset which consists of a total of 163 WSCIs from 27 patients. We combineCNN + PCA + RFmodel andCNN + RFmodel, respectively, with four feature extraction strategies to conduct eight comparative experiments on the 163 WSCIs. Experimental results demonstrate that when using multilevel feature fusion strategy, the classification accuracy of theCNN + PCA + RFmodel reaches the highest to 99.39%. In addition, theCNN + PCA + RFmodel conducting the multilevel feature fusion strategy performs better than that conducting single‐level feature extraction strategies.
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 | 0.001 | 0.001 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 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 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".