Name
Abstract | ||
|---|---|---|
In this paper, we introduced our work in gigapixel crowd counting, the main challenges of this task are about three aspects. First, there is no model or device permitting us to train a network using gigapixel image or do detecting straightly on them. So we segment these images into small pieces in size of 19201200 to train our network. The second problem is that all the images contain thousands of people which form extremely congested scenes. Our solution is using Dilated Convolution Neural Network which showed good performance on understanding highly congested scenes. The third thing is about accuracy of counting. We will do three times of counting in different scales(origin, 16 times small, 256 times small) and calculate the weighted mean of them as final result. |
Year | DOI | Venue |
|---|---|---|
2019 | 10.1109/ICMEW.2019.00079 | 2019 IEEE International Conference on Multimedia & Expo Workshops (ICMEW) |
Keywords | Field | DocType |
CSRNet,Crowd Counting,Crowd Density Estimation,Multi-Scale | Computer vision,Pattern recognition,Computer science,Convolutional neural network,Artificial intelligence,Crowd counting,Gigapixel image | Conference |
ISSN | ISBN | Citations |
2330-7927 | 978-1-5386-9215-8 | 0 |
PageRank | References | Authors |
0.34 | 0 | 4 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Zhijie Cao | 1 | 0 | 0.68 |
| Renyou Yan | 2 | 0 | 0.34 |
| Yiyong Huang | 3 | 0 | 0.34 |
| Zhiru Shi | 4 | 0 | 0.34 |