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.

Block diagram of the proposed ARC-Net framework integrating residual and attention modules for deepfake image detection.

More »

Fig 1 Expand

Table 1.

Dataset split details for DeepFake datasets used in ARC-Net evaluation.

More »

Table 1 Expand

Fig 2.

Sample images of extended dataset:

(a) Real Images and (b) Fake Images.

More »

Fig 2 Expand

Table 2.

Preprocessing and Data Augmentation Techniques with Parameters.

More »

Table 2 Expand

Table 3.

Model Parameter details.

More »

Table 3 Expand

Table 4.

Notation and Symbols Used in Algorithm 1.

More »

Table 4 Expand

Fig 3.

Overall architecture of proposed ARC-Net showing integration of residual and attention blocks atop EfficientNet-B0.

More »

Fig 3 Expand

Fig 4.

Architecture of the Grad-CAM module used for visualizing attention regions with the ARC-Net model.

More »

Fig 4 Expand

Table 5.

Result Evaluation of All Models.

More »

Table 5 Expand

Table 6.

Classification Report of Models.

More »

Table 6 Expand

Fig 5.

ARC-Net performance analysis.

(a) Training/validation accuracy over 30 epochs, and (b) ROC curve showing high AUC score on the ‘140k Real and Fake Faces’ dataset.

More »

Fig 5 Expand

Fig 6.

DenseNet121 performance analysis.

(a) Training/validation accuracy over 30 epochs, and (b) ROC curve on the ‘140k Real and Fake Faces’ dataset.

More »

Fig 6 Expand

Fig 7.

Confusion matrices illustrating the classification performance of ARC-Net.

(a) and DenseNet121 (b) on the ‘140k Real and Fake Faces’ dataset.

More »

Fig 7 Expand

Table 7.

Result Evaluation of ARC-Net model on Deepfake Dataset and Deepfake Database.

More »

Table 7 Expand

Fig 8.

(a) Train validation accuracy plot and (b) Receiver operating characteristic curve of ARC-Net on the Deepfake dataset.

More »

Fig 8 Expand

Fig 9.

(a) Train validation accuracy plot and (b) Receiver operating characteristic curve of ARC-Net on the deepfake database.

More »

Fig 9 Expand

Table 8.

Comparison of classification performance across various research studies on the DeepFake face detection with our proposed ARC-Net model.

More »

Table 8 Expand

Fig 10.

Confusion matrix of ARC-Net in (a) deepfake dataset and (b) deepfake database dataset.

More »

Fig 10 Expand

Fig 11.

Correct and incorrect test examples from the deepfake dataset with ARC-Net.

More »

Fig 11 Expand

Table 9.

Model Configurations and Experimental Setup.

More »

Table 9 Expand

Table 10.

Comparison of Model A and Model B on the Held-Out South Asian Test Set (100 Real Images).

More »

Table 10 Expand

Fig 12.

False Positive Rate change and McNemar contingency table between Model A and Model B.

More »

Fig 12 Expand

Fig 13.

Confidence distributions of mean probability of real for Model A (blue) and Model B (orange).

More »

Fig 13 Expand

Table 11.

Dataset Distribution and Results of OOD Evaluation 01.

More »

Table 11 Expand

Fig 14.

Performance comparison of OOD-01.

More »

Fig 14 Expand

Table 12.

Dataset Distribution and Results of OOD Evaluation 02.

More »

Table 12 Expand

Fig 15.

Performance comparison of OOD-02.

More »

Fig 15 Expand

Table 13.

Experiment Configurations.

More »

Table 13 Expand

Table 14.

Performance of ablation variants and the full ARC-Net model on the hybrid dataset.

More »

Table 14 Expand

Fig 16.

Confusion Matrix of ARC-Net analysis:

(a) Baseline EffecientNetB0 (Experiment A), (b) Attention bassed CNN (Experiment B) and (c) Residual CNN (Experiment C)in hybrid dataset. (d) DenseNet121-based CNN with residual and attention modules (Experiment D).

More »

Fig 16 Expand

Fig 17.

Grad-CAM Explanation and Heatmap with ARC-Net.

More »

Fig 17 Expand

Fig 18.

LIME Explanation for ARC-Net.

More »

Fig 18 Expand

Fig 19.

Grad-CAM Explanation and Heatmap for DenseNet121, ResNet50, and ARC-Net on real human faces.

More »

Fig 19 Expand

Table 15.

Representative IoU, Dice, and Peak-Activation Scores for a Subset of Test Images. The complete dataset comprises 200 images; only a concise selection is included here for clarity. Aggregate metrics (mean and 95% CI) are computed over the full test set.

More »

Table 15 Expand