Hybrid multimodal deepfake detection in medical images using ConvNeXtV2-Tiny with Triplet Attention

Key Points

• Deepfake detection accuracy increased significantly with hybrid methods, enhancing reliability in medical imaging.
• The model achieved a 95% detection rate across various medical image types, including X-rays and MRIs.
• Hybrid multimodal analysis utilized triplet attention to improve feature extraction from medical images.
• Advancements may lead to increased trust in diagnostic tools, vital for patient care and safety.