Name
Abstract | ||
|---|---|---|
Crowd counting has played a substantial role in intelligent surveillance. This work presents a multi-scale multi-task convolutional neural network (MSMT-CNN) to estimate accurate density maps, thus can count the crowd through summing up all values in the estimated density maps. The ground truth density maps used for training are generated by a novel adaptive human-shaped kernel. In addition to resolving the scale problem with the multi-scale strategy, the multi-task learning strategy is added so as to make the estimated density maps more accurate. A weighted loss function is proposed to enhance the activations in dense regions and suppress the background noise. Experimental results on two benchmarking datasets reveal the strong ability of MSMT-CNN. Compared with existing crowd counting methods, the root mean squared error is decreased by 39.8 on the UCF_CC_50 dataset, and the mean absolute error is decreased by 2.3 on the World Expo’10 dataset. Furthermore, the evaluations in practical bus videos verify the practicability of our MSMT-CNN. |
Year | DOI | Venue |
|---|---|---|
2020 | 10.1007/s11042-019-08467-3 | Multimedia Tools and Applications |
Keywords | Field | DocType |
Crowd counting, Multi-task learning, Convolutional neural network, Adaptive human-shaped kernel, Weighted loss function | Kernel (linear algebra),Multi-task learning,Background noise,Pattern recognition,Computer science,Convolutional neural network,Mean squared error,Ground truth,Artificial intelligence,Crowd counting,Benchmarking | Journal |
Volume | Issue | ISSN |
79 | 3-4 | 1573-7721 |
Citations | PageRank | References |
1 | 0.36 | 0 |
Authors | ||
5 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Jinmeng Cao | 1 | 13 | 3.21 |
| Biao Yang | 2 | 10 | 1.83 |
| Nan Wang | 3 | 93 | 27.47 |
| Hai Wang | 4 | 40 | 6.46 |
| Yingfeng Cai | 5 | 104 | 25.57 |