Object Counting
The goal of Object Counting task is to count the number of object instances in a single image or video sequence. It has many real-world applications such as traffic flow monitoring, crowdedness estimation, and product counting.
Source: Learning to Count Objects with Few Exemplar Annotations
Papers
Showing 1–10 of 158 papers
All datasetsFSC147CARPKPascal VOC 2007 count-testCOCO count-testTallyQA-ComplexTallyQA-SimpleHowMany-QAPASCAL VOCOmnicount-191TRANCOS
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | YOLO (2016) | MAE | 156 | — | Unverified |
| 2 | YOLO9000opt (2017) | MAE | 130.4 | — | Unverified |
| 3 | Faster R-CNN (2015) | MAE | 39.88 | — | Unverified |
| 4 | RetinaNet (2018) | MAE | 24.58 | — | Unverified |
| 5 | LPN Counting (2017) | MAE | 22.76 | — | Unverified |
| 6 | One-Look Regression (2016) | MAE | 21.88 | — | Unverified |
| 7 | RetinaNet (2018) | MAE | 16.62 | — | Unverified |
| 8 | CounTX (uses arbitrary text input to specify object to count, used "the cars" for CARPK) | MAE | 8.13 | — | Unverified |
| 9 | Soft-IoU + EM-Merger unit | MAE | 6.77 | — | Unverified |
| 10 | VLCounter | MAE | 6.46 | — | Unverified |