Atrous convolutions spatial pyramid network for crowd counting and density estimation
Scale variation because of perspective distortion is still a challenge for crowd analysis. To address this problem, an atrous convolutions spatial pyramid network (ACSPNet) is proposed to perform crowd counts and density maps for both sparse and congested scenarios. Atrous Convolutions sequenced wit...
Saved in:
Main Authors: | , , |
---|---|
Other Authors: | |
Format: | Article |
Language: | English |
Published: |
2021
|
Subjects: | |
Online Access: | https://hdl.handle.net/10356/151340 |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Nanyang Technological University |
Language: | English |
id |
sg-ntu-dr.10356-151340 |
---|---|
record_format |
dspace |
spelling |
sg-ntu-dr.10356-1513402021-07-09T01:29:27Z Atrous convolutions spatial pyramid network for crowd counting and density estimation Ma, Junjie Dai, Yaping Tan, Yap Peng School of Electrical and Electronic Engineering Engineering::Electrical and electronic engineering Crowd Counting Crowd Density Estimation Scale variation because of perspective distortion is still a challenge for crowd analysis. To address this problem, an atrous convolutions spatial pyramid network (ACSPNet) is proposed to perform crowd counts and density maps for both sparse and congested scenarios. Atrous Convolutions sequenced with increasing atrous rates are utilized to exaggerate the receptive field and maintain the resolution of extracted features. Different rates of atrous convolution blocks in the pyramid are skip-connected to integrate multi-scale information and extent scale perception ability. Atrous Spatial Pyramid Pooling (ASPP) is employed to resample information at different scales and contain global context. We evaluate our ACSPNet on five challenging benchmark crowd counting datasets and our method achieves state-of-the-art mean absolute error (MAE) and mean squared error (MSE) performances. Info-communications Media Development Authority (IMDA) Nanyang Technological University This research was carried out at the Rapid-Rich Object Search (ROSE) Lab at Nanyang Technological University, Singapore. The ROSE Lab is supported by the Infocomm Media Development Authority, Singapore. 2021-07-09T01:29:27Z 2021-07-09T01:29:27Z 2019 Journal Article Ma, J., Dai, Y. & Tan, Y. P. (2019). Atrous convolutions spatial pyramid network for crowd counting and density estimation. Neurocomputing, 350, 91-101. https://dx.doi.org/10.1016/j.neucom.2019.03.065 0925-2312 0000-0002-8593-3074 https://hdl.handle.net/10356/151340 10.1016/j.neucom.2019.03.065 2-s2.0-85064698107 350 91 101 en Neurocomputing © 2019 Elsevier B.V. All rights reserved. |
institution |
Nanyang Technological University |
building |
NTU Library |
continent |
Asia |
country |
Singapore Singapore |
content_provider |
NTU Library |
collection |
DR-NTU |
language |
English |
topic |
Engineering::Electrical and electronic engineering Crowd Counting Crowd Density Estimation |
spellingShingle |
Engineering::Electrical and electronic engineering Crowd Counting Crowd Density Estimation Ma, Junjie Dai, Yaping Tan, Yap Peng Atrous convolutions spatial pyramid network for crowd counting and density estimation |
description |
Scale variation because of perspective distortion is still a challenge for crowd analysis. To address this problem, an atrous convolutions spatial pyramid network (ACSPNet) is proposed to perform crowd counts and density maps for both sparse and congested scenarios. Atrous Convolutions sequenced with increasing atrous rates are utilized to exaggerate the receptive field and maintain the resolution of extracted features. Different rates of atrous convolution blocks in the pyramid are skip-connected to integrate multi-scale information and extent scale perception ability. Atrous Spatial Pyramid Pooling (ASPP) is employed to resample information at different scales and contain global context. We evaluate our ACSPNet on five challenging benchmark crowd counting datasets and our method achieves state-of-the-art mean absolute error (MAE) and mean squared error (MSE) performances. |
author2 |
School of Electrical and Electronic Engineering |
author_facet |
School of Electrical and Electronic Engineering Ma, Junjie Dai, Yaping Tan, Yap Peng |
format |
Article |
author |
Ma, Junjie Dai, Yaping Tan, Yap Peng |
author_sort |
Ma, Junjie |
title |
Atrous convolutions spatial pyramid network for crowd counting and density estimation |
title_short |
Atrous convolutions spatial pyramid network for crowd counting and density estimation |
title_full |
Atrous convolutions spatial pyramid network for crowd counting and density estimation |
title_fullStr |
Atrous convolutions spatial pyramid network for crowd counting and density estimation |
title_full_unstemmed |
Atrous convolutions spatial pyramid network for crowd counting and density estimation |
title_sort |
atrous convolutions spatial pyramid network for crowd counting and density estimation |
publishDate |
2021 |
url |
https://hdl.handle.net/10356/151340 |
_version_ |
1705151283124502528 |