Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

Scheme of the proposed model.

More »

Fig 1 Expand

Table 1.

Recent meta-heuristic optimization in deep learning approaches applied to brain tumor analysis.

More »

Table 1 Expand

Fig 2.

Visualization of the BraTS2023 dataset collection, the range of MRI modalities and ground truth segmentation.

More »

Fig 2 Expand

Fig 3.

Augmented MRI modalities and corresponding segmentation labels.

More »

Fig 3 Expand

Fig 4.

Final preprocessed MRI modalities and corresponding labels.

More »

Fig 4 Expand

Fig 5.

Proposed architecture.

More »

Fig 5 Expand

Fig 6.

Sample image of proposed method results.

More »

Fig 6 Expand

Table 2.

Corresponding sample image metric values.

More »

Table 2 Expand

Table 3.

Quantitative metrics comparison for Sample 100 (validation case).

More »

Table 3 Expand

Fig 7.

Sample image of original and corresponding image for ground truth.

More »

Fig 7 Expand

Fig 8.

Comparison of different algorithms (Sample 100).

More »

Fig 8 Expand

Fig 9.

Comparison of HD95 for different algorithms on Sample 100.

More »

Fig 9 Expand

Table 4.

Quantitative metrics comparison for Sample 77 (validation case).

More »

Table 4 Expand

Fig 10.

Comparison of different algorithms (Sample 77).

More »

Fig 10 Expand

Fig 11.

Comparison of HD95 for different algorithms on Sample 77.

More »

Fig 11 Expand

Fig 12.

Example of no tumor prediction segmentation.

More »

Fig 12 Expand

Fig 13.

Training dynamics of the proposed GA-ResUNetGAN model.

(a) Training loss curve with convergence; (b) Training accuracy showing rapid improvement; (c)-(e) Validation DSC curves for TC, WT, and ET indicating high segmentation performance.

More »

Fig 13 Expand

Table 5.

Comparison of key metrics with and without augmentation (identical training budget in both conditions).

More »

Table 5 Expand

Fig 14.

Error of segmentation because of over fitting results of sample image.

More »

Fig 14 Expand

Table 6.

Accurate hyperparameter settings for GA-ResUNetGAN model.

More »

Table 6 Expand

Table 7.

5-fold cross-validation: Best validation Dice score per fold (mean ± std across folds).

More »

Table 7 Expand

Table 8.

10-Table 9Table 10fold cross-validation: Best validation Dice score per fold (mean ± std across folds).

More »

Table 8 Expand

Table 9.

Ablation of architectural components and their effect on segmentation performance.

More »

Table 9 Expand

Table 10.

Wilcoxon signed-rank test comparing GA-ResUNetGAN with baseline models on BraTS2023 validation (n = 22).

More »

Table 10 Expand

Table 11.

Model complexity: trainable parameters, FLOPs per 96 × 96 × 96 patch, and inference time per full MRI volume.

More »

Table 11 Expand

Fig 15.

Comparison of CPU time, system time, and total execution time (in seconds) for different segmentation models over two training epochs.

Average runtime per epoch is annotated above each bar.

More »

Fig 15 Expand

Table 12.

Comparison of Brain Tumor Segmentation Models on BraTS Datasets.

More »

Table 12 Expand