Manoj Acharya, Kushal Kafle, Christopher Kanan
Most counting questions in visual question answering (VQA) datasets are simple and require no more than object detection. Here, we study algorithms for complex counting questions that involve relationships between objects, attribute identification, reasoning, and more. To do this, we created TallyQA, the world's largest dataset for open-ended counting. We propose a new algorithm for counting that uses relation networks with region proposals. Our method lets relation networks be efficiently used with high-resolution imagery. It yields state-of-the-art results compared to baseline and recent systems on both TallyQA and the HowMany-QA benchmark.
| Task | Dataset | Metric | Value | Model |
|---|---|---|---|---|
| Object Counting | HowMany-QA | Accuracy | 60.3 | RCN |
| Object Counting | HowMany-QA | RMSE | 2.35 | RCN |
| Object Counting | TallyQA-Complex | Accuracy | 56.2 | RCN |
| Object Counting | TallyQA-Complex | RMSE | 1.43 | RCN |
| Object Counting | TallyQA-Simple | Accuracy | 71.8 | RCN |
| Object Counting | TallyQA-Simple | RMSE | 1.13 | RCN |