NCT: Noise-Control Tracker
Video not available.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Rendering of new sequences is currently deactivated due to heavy load.
Benchmark:
MOT17 |
Short name:
NCT
Detector:
Public
Description:
The experiments were conducted on a server equipped with a Tesla V100 GPU, and the operating system is Ubuntu 20.04. Our implementation is based on CenterNet . Deep Layer Aggregation (DLA) is taken as the network backbone, and it is optimized with Adam with a learning rate of 1.16e-4 and a batch size of 3. Data augmentations include random horizontal flipping, random resized cropping, and color jittering. For all experiments, the networks are trained for 150 epochs. The learning rate is dropped by a factor of 10 at the 60th epoch.
Reference:
K. Zeng, Y. You, T. Shen, Q. Wang, Z. Tao, Z. Wang, Q. Liu. NCT:noise-control multi-object tracking. In Complex & Intelligent Systems, 2023.
Last submitted:
May 10, 2023 (11 months ago)
Published:
May 10, 2023 at 15:32:07 CET
Submissions:
1
Project page / code:
Open source:
Yes
Hardware:
V100
Runtime:
13.9 Hz
Benchmark performance:
| Sequence | MOTA | IDF1 | HOTA | MT | ML | FP | FN | Rcll | Prcn | AssA | DetA | AssRe | AssPr | DetRe | DetPr | LocA | FAF | ID Sw. | Frag |
| MOT17 | 69.5 | 68.5 | 54.6 | 1,092 (46.4) | 399 (16.9) | 65,463 | 101,471 | 82.0 | 87.6 | 53.7 | 56.1 | 59.6 | 71.5 | 64.7 | 69.1 | 79.9 | 3.7 | 4,919 (0.0) | 4,054 (0.0) |
Detailed performance:
| Sequence | MOTA | IDF1 | HOTA | MT | ML | FP | FN | Rcll | Prcn | AssA | DetA | AssRe | AssPr | DetRe | DetPr | LocA | FAF | ID Sw. | Frag |
| MOT17-01-DPM | 52.0 | 52.8 | 43.3 | 13 | 5 | 1,030 | 2,033 | 68.5 | 81.1 | 41.1 | 46.0 | 46.4 | 66.3 | 54.1 | 64.1 | 78.4 | 2.3 | 34 | 44 |
| MOT17-01-FRCNN | 53.3 | 50.7 | 41.4 | 13 | 5 | 989 | 1,990 | 69.1 | 81.8 | 37.6 | 46.0 | 43.1 | 62.8 | 54.1 | 64.1 | 78.3 | 2.2 | 36 | 43 |
| MOT17-01-SDP | 52.0 | 52.8 | 43.3 | 13 | 5 | 1,030 | 2,033 | 68.5 | 81.1 | 41.1 | 46.0 | 46.4 | 66.3 | 54.1 | 64.1 | 78.4 | 2.3 | 34 | 44 |
| MOT17-03-DPM | 84.9 | 80.4 | 63.1 | 136 | 1 | 10,992 | 4,737 | 95.5 | 90.1 | 61.5 | 65.2 | 66.7 | 75.9 | 75.0 | 70.8 | 80.2 | 7.3 | 108 | 190 |
| MOT17-03-FRCNN | 84.9 | 80.3 | 63.1 | 136 | 1 | 11,001 | 4,745 | 95.5 | 90.1 | 61.5 | 65.2 | 66.7 | 75.9 | 75.0 | 70.8 | 80.2 | 7.3 | 108 | 190 |
| MOT17-03-SDP | 84.9 | 80.3 | 63.1 | 136 | 1 | 11,003 | 4,747 | 95.5 | 90.1 | 61.5 | 65.2 | 66.7 | 75.9 | 75.0 | 70.8 | 80.2 | 7.3 | 108 | 189 |
| MOT17-06-DPM | 52.7 | 52.0 | 40.1 | 92 | 49 | 1,816 | 3,486 | 70.4 | 82.0 | 34.9 | 46.3 | 50.0 | 50.6 | 54.6 | 63.6 | 78.0 | 1.5 | 273 | 228 |
| MOT17-06-FRCNN | 52.7 | 52.0 | 40.1 | 92 | 49 | 1,816 | 3,486 | 70.4 | 82.0 | 34.9 | 46.3 | 50.0 | 50.6 | 54.6 | 63.6 | 78.0 | 1.5 | 273 | 228 |
| MOT17-06-SDP | 52.7 | 52.0 | 40.1 | 92 | 49 | 1,816 | 3,486 | 70.4 | 82.0 | 34.9 | 46.3 | 50.0 | 50.6 | 54.6 | 63.6 | 78.0 | 1.5 | 273 | 228 |
| MOT17-07-DPM | 57.8 | 50.7 | 42.1 | 25 | 6 | 2,515 | 4,466 | 73.6 | 83.2 | 37.5 | 48.7 | 42.0 | 66.0 | 57.3 | 64.8 | 78.4 | 5.0 | 148 | 217 |
| MOT17-07-FRCNN | 57.8 | 50.7 | 42.1 | 25 | 6 | 2,515 | 4,466 | 73.6 | 83.2 | 37.5 | 48.7 | 42.0 | 66.0 | 57.3 | 64.8 | 78.4 | 5.0 | 148 | 217 |
| MOT17-07-SDP | 57.8 | 50.7 | 42.1 | 25 | 6 | 2,515 | 4,466 | 73.6 | 83.2 | 37.5 | 48.7 | 42.0 | 66.0 | 57.3 | 64.8 | 78.4 | 5.0 | 148 | 217 |
| MOT17-08-DPM | 52.2 | 45.1 | 39.2 | 25 | 9 | 1,440 | 8,438 | 60.1 | 89.8 | 35.2 | 45.0 | 42.3 | 60.6 | 49.2 | 73.5 | 81.2 | 2.3 | 224 | 249 |
| MOT17-08-FRCNN | 52.2 | 45.1 | 39.2 | 25 | 9 | 1,440 | 8,438 | 60.1 | 89.8 | 35.2 | 45.0 | 42.3 | 60.6 | 49.2 | 73.5 | 81.2 | 2.3 | 224 | 249 |
| MOT17-08-SDP | 52.2 | 45.1 | 39.2 | 25 | 9 | 1,440 | 8,438 | 60.1 | 89.8 | 35.2 | 45.0 | 42.3 | 60.6 | 49.2 | 73.5 | 81.2 | 2.3 | 224 | 249 |
| MOT17-12-DPM | 56.3 | 67.5 | 52.4 | 35 | 24 | 1,006 | 2,739 | 68.4 | 85.5 | 57.3 | 48.1 | 64.0 | 74.2 | 55.3 | 69.1 | 81.4 | 1.1 | 40 | 57 |
| MOT17-12-FRCNN | 56.3 | 67.5 | 52.4 | 35 | 24 | 1,006 | 2,739 | 68.4 | 85.5 | 57.3 | 48.1 | 64.0 | 74.2 | 55.3 | 69.1 | 81.4 | 1.1 | 40 | 57 |
| MOT17-12-SDP | 56.3 | 67.5 | 52.4 | 35 | 24 | 1,006 | 2,739 | 68.4 | 85.5 | 57.3 | 48.1 | 64.0 | 74.2 | 55.3 | 69.1 | 81.4 | 1.1 | 40 | 57 |
| MOT17-14-DPM | 36.3 | 50.4 | 37.7 | 38 | 39 | 3,029 | 7,933 | 57.1 | 77.7 | 37.7 | 38.1 | 42.4 | 63.8 | 45.0 | 61.2 | 78.0 | 4.0 | 812 | 367 |
| MOT17-14-FRCNN | 36.3 | 50.4 | 37.7 | 38 | 39 | 3,029 | 7,933 | 57.1 | 77.7 | 37.7 | 38.1 | 42.4 | 63.8 | 45.0 | 61.2 | 78.0 | 4.0 | 812 | 367 |
| MOT17-14-SDP | 36.3 | 50.4 | 37.7 | 38 | 39 | 3,029 | 7,933 | 57.1 | 77.7 | 37.7 | 38.1 | 42.4 | 63.8 | 45.0 | 61.2 | 78.0 | 4.0 | 812 | 367 |
Raw data:
