Redesigned Skip-Network for Crowd Counting with Dilated Convolution and Backward Connection
Autor: | Toshiaki Kondo, Atsuo Yoshitaka, Pished Bunnun, Sorn Sooksatra |
---|---|
Jazyk: | angličtina |
Rok vydání: | 2020 |
Předmět: |
Mean squared error
Computer science Mean absolute error 02 engineering and technology 010501 environmental sciences lcsh:Computer applications to medicine. Medical informatics 01 natural sciences surveillance system Article lcsh:QA75.5-76.95 Robustness (computer science) 0202 electrical engineering electronic engineering information engineering crowd counting Radiology Nuclear Medicine and imaging lcsh:Photography Electrical and Electronic Engineering Semantic information Crowd counting 0105 earth and related environmental sciences business.industry Pattern recognition dilated convolution lcsh:TR1-1050 Computer Graphics and Computer-Aided Design skip connection regression-based approach lcsh:R858-859.7 020201 artificial intelligence & image processing Computer Vision and Pattern Recognition Artificial intelligence lcsh:Electronic computers. Computer science Crowd density business |
Zdroj: | Journal of Imaging, Vol 6, Iss 28, p 28 (2020) Journal of Imaging Volume 6 Issue 5 |
Popis: | Crowd counting is a challenging task dealing with the variation of an object scale and a crowd density. Existing works have emphasized on skip connections by integrating shallower layers with deeper layers, where each layer extracts features in a different object scale and crowd density. However, only high-level features are emphasized while ignoring low-level features. This paper proposes an estimation network by passing high-level features to shallow layers and emphasizing its low-level feature. Since an estimation network is a hierarchical network, a high-level feature is also emphasized by an improved low-level feature. Our estimation network consists of two identical networks for extracting a high-level feature and estimating the final result. To preserve semantic information, dilated convolution is employed without resizing the feature map. Our method was tested in three datasets for counting humans and vehicles in a crowd image. The counting performance is evaluated by mean absolute error and root mean squared error indicating the accuracy and robustness of an estimation network, respectively. The experimental result shows that our network outperforms other related works in a high crowd density and is effective for reducing over-counting error in the overall case. |
Databáze: | OpenAIRE |
Externí odkaz: |
Pro tento záznam nejsou dostupné žádné jednotky.