Matches in SemOpenAlex for { <https://semopenalex.org/work/W4381736354> ?p ?o ?g. }
Showing items 1 to 82 of
82
with 100 items per page.
- W4381736354 abstract "To validate the effectiveness of an approach called batch-balanced focal loss (BBFL) in enhancing convolutional neural network (CNN) classification performance on imbalanced datasets.BBFL combines two strategies to tackle class imbalance: (1) batch-balancing to equalize model learning of class samples and (2) focal loss to add hard-sample importance to the learning gradient. BBFL was validated on two imbalanced fundus image datasets: a binary retinal nerve fiber layer defect (RNFLD) dataset (n=7,258) and a multiclass glaucoma dataset (n=7,873). BBFL was compared to several imbalanced learning techniques, including random oversampling (ROS), cost-sensitive learning, and thresholding, based on three state-of-the-art CNNs. Accuracy, F1-score, and the area under the receiver operator characteristic curve (AUC) were used as the performance metrics for binary classification. Mean accuracy and mean F1-score were used for multiclass classification. Confusion matrices, t-distributed neighbor embedding plots, and GradCAM were used for the visual assessment of performance.In binary classification of RNFLD, BBFL with InceptionV3 (93.0% accuracy, 84.7% F1, 0.971 AUC) outperformed ROS (92.6% accuracy, 83.7% F1, 0.964 AUC), cost-sensitive learning (92.5% accuracy, 83.8% F1, 0.962 AUC), and thresholding (91.9% accuracy, 83.0% F1, 0.962 AUC) and others. In multiclass classification of glaucoma, BBFL with MobileNetV2 (79.7% accuracy, 69.6% average F1 score) outperformed ROS (76.8% accuracy, 64.7% F1), cost-sensitive learning (78.3% accuracy, 67.8.8% F1), and random undersampling (76.5% accuracy, 66.5% F1).The BBFL-based learning method can improve the performance of a CNN model in both binary and multiclass disease classification when the data are imbalanced." @default.
- W4381736354 created "2023-06-24" @default.
- W4381736354 creator A5033429874 @default.
- W4381736354 creator A5034125372 @default.
- W4381736354 creator A5041650010 @default.
- W4381736354 creator A5048812493 @default.
- W4381736354 creator A5049578096 @default.
- W4381736354 creator A5065866225 @default.
- W4381736354 creator A5074284003 @default.
- W4381736354 date "2023-06-23" @default.
- W4381736354 modified "2023-09-27" @default.
- W4381736354 title "Batch-balanced focal loss: a hybrid solution to class imbalance in deep learning" @default.
- W4381736354 cites W2076063813 @default.
- W4381736354 cites W2884561390 @default.
- W4381736354 cites W3005214817 @default.
- W4381736354 cites W3072781859 @default.
- W4381736354 doi "https://doi.org/10.1117/1.jmi.10.5.051809" @default.
- W4381736354 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/37361550" @default.
- W4381736354 hasPublicationYear "2023" @default.
- W4381736354 type Work @default.
- W4381736354 citedByCount "0" @default.
- W4381736354 crossrefType "journal-article" @default.
- W4381736354 hasAuthorship W4381736354A5033429874 @default.
- W4381736354 hasAuthorship W4381736354A5034125372 @default.
- W4381736354 hasAuthorship W4381736354A5041650010 @default.
- W4381736354 hasAuthorship W4381736354A5048812493 @default.
- W4381736354 hasAuthorship W4381736354A5049578096 @default.
- W4381736354 hasAuthorship W4381736354A5065866225 @default.
- W4381736354 hasAuthorship W4381736354A5074284003 @default.
- W4381736354 hasConcept C115961682 @default.
- W4381736354 hasConcept C119857082 @default.
- W4381736354 hasConcept C12267149 @default.
- W4381736354 hasConcept C123860398 @default.
- W4381736354 hasConcept C126322002 @default.
- W4381736354 hasConcept C136536468 @default.
- W4381736354 hasConcept C153180895 @default.
- W4381736354 hasConcept C154945302 @default.
- W4381736354 hasConcept C169258074 @default.
- W4381736354 hasConcept C191178318 @default.
- W4381736354 hasConcept C41008148 @default.
- W4381736354 hasConcept C44249647 @default.
- W4381736354 hasConcept C58471807 @default.
- W4381736354 hasConcept C66905080 @default.
- W4381736354 hasConcept C68443243 @default.
- W4381736354 hasConcept C71924100 @default.
- W4381736354 hasConcept C81363708 @default.
- W4381736354 hasConceptScore W4381736354C115961682 @default.
- W4381736354 hasConceptScore W4381736354C119857082 @default.
- W4381736354 hasConceptScore W4381736354C12267149 @default.
- W4381736354 hasConceptScore W4381736354C123860398 @default.
- W4381736354 hasConceptScore W4381736354C126322002 @default.
- W4381736354 hasConceptScore W4381736354C136536468 @default.
- W4381736354 hasConceptScore W4381736354C153180895 @default.
- W4381736354 hasConceptScore W4381736354C154945302 @default.
- W4381736354 hasConceptScore W4381736354C169258074 @default.
- W4381736354 hasConceptScore W4381736354C191178318 @default.
- W4381736354 hasConceptScore W4381736354C41008148 @default.
- W4381736354 hasConceptScore W4381736354C44249647 @default.
- W4381736354 hasConceptScore W4381736354C58471807 @default.
- W4381736354 hasConceptScore W4381736354C66905080 @default.
- W4381736354 hasConceptScore W4381736354C68443243 @default.
- W4381736354 hasConceptScore W4381736354C71924100 @default.
- W4381736354 hasConceptScore W4381736354C81363708 @default.
- W4381736354 hasIssue "05" @default.
- W4381736354 hasLocation W43817363541 @default.
- W4381736354 hasLocation W43817363542 @default.
- W4381736354 hasOpenAccess W4381736354 @default.
- W4381736354 hasPrimaryLocation W43817363541 @default.
- W4381736354 hasRelatedWork W2987667774 @default.
- W4381736354 hasRelatedWork W3176807344 @default.
- W4381736354 hasRelatedWork W3215867059 @default.
- W4381736354 hasRelatedWork W4221077026 @default.
- W4381736354 hasRelatedWork W4285133259 @default.
- W4381736354 hasRelatedWork W4285281467 @default.
- W4381736354 hasRelatedWork W4291692947 @default.
- W4381736354 hasRelatedWork W4309224780 @default.
- W4381736354 hasRelatedWork W4322008322 @default.
- W4381736354 hasRelatedWork W4384300587 @default.
- W4381736354 hasVolume "10" @default.
- W4381736354 isParatext "false" @default.
- W4381736354 isRetracted "false" @default.
- W4381736354 workType "article" @default.