Table 3 The performance of the proposed model, the KLT tracker, and the BM tracker. (TSR: Tracking Success Rate)
From: A deep learning approach to track Arabidopsis seedlings’ circumnutation from time-lapse videos
Video | Proposed method | KLT tracker | BM tracker | ||||||
|---|---|---|---|---|---|---|---|---|---|
TSR | Error mean (mm) | Error std (mm) | TSR | Error mean (mm) | Error std (mm) | TSR | Error mean (mm) | Error std (mm) | |
Val_1 | 100% | 0.32 | 0.19 | 100% | 1.16 | 0.28 | 0.0% | N/A | N/A |
Val_2 | 100% | 1.52 | 0.72 | 34.8% | 1.84 | 0.80 | 70.4% | 4.17 | 0.75 |
Test_1 | 100% | 0.91 | 0.48 | 100% | 1.13 | 0.65 | 100% | 2.46 | 0.64 |
Test_2 | 100% | 0.94 | 0.53 | 98.0% | 1.01 | 0.44 | 42.5% | 1.04 | 0.41 |
Test_3 | 100% | 1.43 | 0.52 | 100% | 2.10 | 0.93 | 25.2% | 3.19 | 0.59 |
Average | 100% | 1.02 | 0.49 | 86.6% | 1.45 | 0.62 | 47.6% | 2.72 | 0.60 |