Crowd counting with fully convolutional neural network Conference Paper
| Authors: | Liu, M.; Jiang, J.; Guo, Z.; Wang, Z.; Liu, Y. |
| Title: | Crowd counting with fully convolutional neural network |
| Conference Title: | 25th IEEE International Conference on Image Processing (ICIP 2018) |
| Abstract: | Crowd counting estimation is an extremely challenging task due to various crowded scenarios. In this paper, we present a deep learning framework for crowd counting from a single static image with different number of people and arbitrary perspective. In the design of convolutional neural network structure, we employ the VGG16 model but drop the fully connected layers. Meanwhile, high-level features are combined with low-level features through laterally connected feature pyramid network by element-wise addition to ensure higher resolution and more context information. Extensive experiments are conducted on ShanghaiTech and UCFCC50 datasets. The results show that our model achieves the lowest mean absolute error (MAE) and comparable mean square error (MSE), and outperforms the current state-of-the-art methods. © 2018 IEEE. |
| Keywords: | image processing; neural networks; learning frameworks; mean square error; deep learning; convolution; convolutional neural network; high-level features; state-of-the-art methods; crowd counting; feature pyramid network; context information; feature pyramid; mean absolute error |
| Journal Title | International Conference on Image Processing. Proceedings |
| Conference Dates: | 2018 Oct 7-10 |
| Conference Location: | Athens, Greece |
| ISBN: | 1522-4880 |
| Publisher: | IEEE |
| Date Published: | 2018-01-01 |
| Start Page: | 953 |
| End Page: | 957 |
| Language: | English |
| DOI: | 10.1109/ICIP.2018.8451787 |
| PROVIDER: | scopus |
| DOI/URL: | |
| Notes: | (ISBN: 978-1-4799-7061-2) -- WQ.L2.4 -- Source: Scopus |
