Table 8 The experimental results of the proposed BPI-MVQA on the VQA-RAD dataset
From: BPI-MVQA: a bi-branch model for medical visual question answering
Type | Accuracy | WBSS | Recall | Precision |
|---|---|---|---|---|
Modality | 0.714 | 0.751 | 0.330 | 0.295 |
Plane | 0.654 | 0.696 | 0.328 | 0.254 |
Organ | 0.60 | 0.627 | 0.467 | 0.417 |
Abnormality | 0.95 | 1.0 | 1.0 | 1.0 |
Pos | 0.656 | 0.727 | 0.481 | 0.460 |
Color | 1.0 | 1.0 | 1.0 | 1.0 |
Size | 0.804 | 0.813 | 0.841 | 0.800 |
Count | 1.0 | 1.0 | 1.0 | 1.0 |
Attribute | 0.95 | 1.0 | 1.0 | 1.0 |
Other | 0.625 | 0.684 | 0.112 | 0.122 |
Pres | 0.811 | 0.82 | 0.538 | 0.536 |
Open-ended | 0.625 | 0.624 | 0.391 | 0.370 |
Close-ended | 0.841 | 0.848 | 0.841 | 0.840 |
All | 0.727 | 0.753 | 0.397 | 0.376 |