Robust Motion Segmentation from Pairwise Matches
About
In this paper we address a classification problem that has not been considered before, namely motion segmentation given pairwise matches only. Our contribution to this unexplored task is a novel formulation of motion segmentation as a two-step process. First, motion segmentation is performed on image pairs independently. Secondly, we combine independent pairwise segmentation results in a robust way into the final globally consistent segmentation. Our approach is inspired by the success of averaging methods. We demonstrate in simulated as well as in real experiments that our method is very effective in reducing the errors in the pairwise motion segmentation and can cope with large number of mismatches.
Federica Arrigoni, Tomas Pajdla• 2019
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Motion Segmentation | Hopkins 155 3-motion sequences | Mean Clustering Error (%)2.67 | 45 | |
| Motion Segmentation | Hopkins 155 (all sequences) | Mean Clustering Error1.37 | 45 | |
| Motion Segmentation | Hopkins 155 2-motion sequences | Classification Error0.01 | 31 | |
| Motion Segmentation | Hopkins 12 | Avg Classification Error4.33 | 20 | |
| Motion Segmentation | Indoor Dataset Penguin sequence 1.0 (test) | Misclassification Error0.0076 | 6 | |
| Motion Segmentation | Indoor Dataset Flowers sequence 1.0 (test) | Error (%)1.23 | 6 | |
| Motion Segmentation | Indoor Dataset Bag sequence 1.0 (test) | Misclassification Error (%)1.52 | 6 | |
| Motion Segmentation | Indoor Dataset Pencils sequence 1.0 (test) | Misclassification Error0.038 | 6 | |
| Motion Segmentation | Indoor Dataset Bears sequence 1.0 (test) | Error Rate4.82 | 6 | |
| Motion Segmentation | helicopter | Misclassification Error (%)2.01 | 3 |
Showing 10 of 13 rows