Four-way Classification (NC, SMC, MCI, AD) on ADNI Image-Gene (test)
95.58AccuracyR-GenIMA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| R-GenIMABackbone=Qwen2025.12 | 95.58 | 90.42 | |
| BERT+RiTBackbone=BERT & RiT, Fusion strategy=simple feature-concatenation2025.12 | 95.09 | 89.97 | |
| R-GenIMABackbone=Llama2025.12 | 95.09 | 90.37 | |
| BERT+Med3DBackbone=BERT & Med3D, Fusion strategy=simple feature-concatenation2025.12 | 94.59 | 87.16 | |
| Qwen+Med3DBackbone=Qwen & Med3D2025.12 | 93.61 | 85.87 | |
| Llama+Med3DBackbone=Llama & Med3D2025.12 | 93.36 | 85.71 | |
| BERT+RiTBackbone=BERT & RiT, Fusion strategy=cross-modal attention–based fusion2025.12 | 92.87 | 85.67 | |
| BERT+Med3DBackbone=BERT & Med3D, Fusion strategy=cross-modal attention–based fusion2025.12 | 90.42 | 82.45 |