Crowd density estimation based on classification activation map and patch density level
Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd count...
Ausführliche Beschreibung
Autor*in: |
Zhu, Liping [verfasserIn] |
---|
Format: |
Artikel |
---|---|
Sprache: |
Englisch |
Erschienen: |
2019 |
---|
Schlagwörter: |
---|
Anmerkung: |
© Springer-Verlag London Ltd., part of Springer Nature 2019 |
---|
Übergeordnetes Werk: |
Enthalten in: Neural computing & applications - Springer London, 1993, 32(2019), 9 vom: 03. Jan., Seite 5105-5116 |
---|---|
Übergeordnetes Werk: |
volume:32 ; year:2019 ; number:9 ; day:03 ; month:01 ; pages:5105-5116 |
Links: |
---|
DOI / URN: |
10.1007/s00521-018-3954-7 |
---|
Katalog-ID: |
OLC2025619855 |
---|
LEADER | 01000caa a22002652 4500 | ||
---|---|---|---|
001 | OLC2025619855 | ||
003 | DE-627 | ||
005 | 20230504132905.0 | ||
007 | tu | ||
008 | 200820s2019 xx ||||| 00| ||eng c | ||
024 | 7 | |a 10.1007/s00521-018-3954-7 |2 doi | |
035 | |a (DE-627)OLC2025619855 | ||
035 | |a (DE-He213)s00521-018-3954-7-p | ||
040 | |a DE-627 |b ger |c DE-627 |e rakwb | ||
041 | |a eng | ||
082 | 0 | 4 | |a 004 |q VZ |
100 | 1 | |a Zhu, Liping |e verfasserin |4 aut | |
245 | 1 | 0 | |a Crowd density estimation based on classification activation map and patch density level |
264 | 1 | |c 2019 | |
336 | |a Text |b txt |2 rdacontent | ||
337 | |a ohne Hilfsmittel zu benutzen |b n |2 rdamedia | ||
338 | |a Band |b nc |2 rdacarrier | ||
500 | |a © Springer-Verlag London Ltd., part of Springer Nature 2019 | ||
520 | |a Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods. | ||
650 | 4 | |a Crowd density estimation | |
650 | 4 | |a Image patch | |
650 | 4 | |a Density level | |
650 | 4 | |a Attention mechanism | |
650 | 4 | |a Classification activation map | |
700 | 1 | |a Li, Chengyang |0 (orcid)0000-0002-1379-1222 |4 aut | |
700 | 1 | |a Yang, Zhongguo |4 aut | |
700 | 1 | |a Yuan, Kun |4 aut | |
700 | 1 | |a Wang, Shang |4 aut | |
773 | 0 | 8 | |i Enthalten in |t Neural computing & applications |d Springer London, 1993 |g 32(2019), 9 vom: 03. Jan., Seite 5105-5116 |w (DE-627)165669608 |w (DE-600)1136944-9 |w (DE-576)032873050 |x 0941-0643 |7 nnns |
773 | 1 | 8 | |g volume:32 |g year:2019 |g number:9 |g day:03 |g month:01 |g pages:5105-5116 |
856 | 4 | 1 | |u https://doi.org/10.1007/s00521-018-3954-7 |z lizenzpflichtig |3 Volltext |
912 | |a GBV_USEFLAG_A | ||
912 | |a SYSFLAG_A | ||
912 | |a GBV_OLC | ||
912 | |a SSG-OLC-MAT | ||
912 | |a GBV_ILN_70 | ||
912 | |a GBV_ILN_2018 | ||
912 | |a GBV_ILN_4277 | ||
951 | |a AR | ||
952 | |d 32 |j 2019 |e 9 |b 03 |c 01 |h 5105-5116 |
author_variant |
l z lz c l cl z y zy k y ky s w sw |
---|---|
matchkey_str |
article:09410643:2019----::rwdniysiainaeocasfctoatvtomp |
hierarchy_sort_str |
2019 |
publishDate |
2019 |
allfields |
10.1007/s00521-018-3954-7 doi (DE-627)OLC2025619855 (DE-He213)s00521-018-3954-7-p DE-627 ger DE-627 rakwb eng 004 VZ Zhu, Liping verfasserin aut Crowd density estimation based on classification activation map and patch density level 2019 Text txt rdacontent ohne Hilfsmittel zu benutzen n rdamedia Band nc rdacarrier © Springer-Verlag London Ltd., part of Springer Nature 2019 Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods. Crowd density estimation Image patch Density level Attention mechanism Classification activation map Li, Chengyang (orcid)0000-0002-1379-1222 aut Yang, Zhongguo aut Yuan, Kun aut Wang, Shang aut Enthalten in Neural computing & applications Springer London, 1993 32(2019), 9 vom: 03. Jan., Seite 5105-5116 (DE-627)165669608 (DE-600)1136944-9 (DE-576)032873050 0941-0643 nnns volume:32 year:2019 number:9 day:03 month:01 pages:5105-5116 https://doi.org/10.1007/s00521-018-3954-7 lizenzpflichtig Volltext GBV_USEFLAG_A SYSFLAG_A GBV_OLC SSG-OLC-MAT GBV_ILN_70 GBV_ILN_2018 GBV_ILN_4277 AR 32 2019 9 03 01 5105-5116 |
spelling |
10.1007/s00521-018-3954-7 doi (DE-627)OLC2025619855 (DE-He213)s00521-018-3954-7-p DE-627 ger DE-627 rakwb eng 004 VZ Zhu, Liping verfasserin aut Crowd density estimation based on classification activation map and patch density level 2019 Text txt rdacontent ohne Hilfsmittel zu benutzen n rdamedia Band nc rdacarrier © Springer-Verlag London Ltd., part of Springer Nature 2019 Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods. Crowd density estimation Image patch Density level Attention mechanism Classification activation map Li, Chengyang (orcid)0000-0002-1379-1222 aut Yang, Zhongguo aut Yuan, Kun aut Wang, Shang aut Enthalten in Neural computing & applications Springer London, 1993 32(2019), 9 vom: 03. Jan., Seite 5105-5116 (DE-627)165669608 (DE-600)1136944-9 (DE-576)032873050 0941-0643 nnns volume:32 year:2019 number:9 day:03 month:01 pages:5105-5116 https://doi.org/10.1007/s00521-018-3954-7 lizenzpflichtig Volltext GBV_USEFLAG_A SYSFLAG_A GBV_OLC SSG-OLC-MAT GBV_ILN_70 GBV_ILN_2018 GBV_ILN_4277 AR 32 2019 9 03 01 5105-5116 |
allfields_unstemmed |
10.1007/s00521-018-3954-7 doi (DE-627)OLC2025619855 (DE-He213)s00521-018-3954-7-p DE-627 ger DE-627 rakwb eng 004 VZ Zhu, Liping verfasserin aut Crowd density estimation based on classification activation map and patch density level 2019 Text txt rdacontent ohne Hilfsmittel zu benutzen n rdamedia Band nc rdacarrier © Springer-Verlag London Ltd., part of Springer Nature 2019 Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods. Crowd density estimation Image patch Density level Attention mechanism Classification activation map Li, Chengyang (orcid)0000-0002-1379-1222 aut Yang, Zhongguo aut Yuan, Kun aut Wang, Shang aut Enthalten in Neural computing & applications Springer London, 1993 32(2019), 9 vom: 03. Jan., Seite 5105-5116 (DE-627)165669608 (DE-600)1136944-9 (DE-576)032873050 0941-0643 nnns volume:32 year:2019 number:9 day:03 month:01 pages:5105-5116 https://doi.org/10.1007/s00521-018-3954-7 lizenzpflichtig Volltext GBV_USEFLAG_A SYSFLAG_A GBV_OLC SSG-OLC-MAT GBV_ILN_70 GBV_ILN_2018 GBV_ILN_4277 AR 32 2019 9 03 01 5105-5116 |
allfieldsGer |
10.1007/s00521-018-3954-7 doi (DE-627)OLC2025619855 (DE-He213)s00521-018-3954-7-p DE-627 ger DE-627 rakwb eng 004 VZ Zhu, Liping verfasserin aut Crowd density estimation based on classification activation map and patch density level 2019 Text txt rdacontent ohne Hilfsmittel zu benutzen n rdamedia Band nc rdacarrier © Springer-Verlag London Ltd., part of Springer Nature 2019 Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods. Crowd density estimation Image patch Density level Attention mechanism Classification activation map Li, Chengyang (orcid)0000-0002-1379-1222 aut Yang, Zhongguo aut Yuan, Kun aut Wang, Shang aut Enthalten in Neural computing & applications Springer London, 1993 32(2019), 9 vom: 03. Jan., Seite 5105-5116 (DE-627)165669608 (DE-600)1136944-9 (DE-576)032873050 0941-0643 nnns volume:32 year:2019 number:9 day:03 month:01 pages:5105-5116 https://doi.org/10.1007/s00521-018-3954-7 lizenzpflichtig Volltext GBV_USEFLAG_A SYSFLAG_A GBV_OLC SSG-OLC-MAT GBV_ILN_70 GBV_ILN_2018 GBV_ILN_4277 AR 32 2019 9 03 01 5105-5116 |
allfieldsSound |
10.1007/s00521-018-3954-7 doi (DE-627)OLC2025619855 (DE-He213)s00521-018-3954-7-p DE-627 ger DE-627 rakwb eng 004 VZ Zhu, Liping verfasserin aut Crowd density estimation based on classification activation map and patch density level 2019 Text txt rdacontent ohne Hilfsmittel zu benutzen n rdamedia Band nc rdacarrier © Springer-Verlag London Ltd., part of Springer Nature 2019 Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods. Crowd density estimation Image patch Density level Attention mechanism Classification activation map Li, Chengyang (orcid)0000-0002-1379-1222 aut Yang, Zhongguo aut Yuan, Kun aut Wang, Shang aut Enthalten in Neural computing & applications Springer London, 1993 32(2019), 9 vom: 03. Jan., Seite 5105-5116 (DE-627)165669608 (DE-600)1136944-9 (DE-576)032873050 0941-0643 nnns volume:32 year:2019 number:9 day:03 month:01 pages:5105-5116 https://doi.org/10.1007/s00521-018-3954-7 lizenzpflichtig Volltext GBV_USEFLAG_A SYSFLAG_A GBV_OLC SSG-OLC-MAT GBV_ILN_70 GBV_ILN_2018 GBV_ILN_4277 AR 32 2019 9 03 01 5105-5116 |
language |
English |
source |
Enthalten in Neural computing & applications 32(2019), 9 vom: 03. Jan., Seite 5105-5116 volume:32 year:2019 number:9 day:03 month:01 pages:5105-5116 |
sourceStr |
Enthalten in Neural computing & applications 32(2019), 9 vom: 03. Jan., Seite 5105-5116 volume:32 year:2019 number:9 day:03 month:01 pages:5105-5116 |
format_phy_str_mv |
Article |
institution |
findex.gbv.de |
topic_facet |
Crowd density estimation Image patch Density level Attention mechanism Classification activation map |
dewey-raw |
004 |
isfreeaccess_bool |
false |
container_title |
Neural computing & applications |
authorswithroles_txt_mv |
Zhu, Liping @@aut@@ Li, Chengyang @@aut@@ Yang, Zhongguo @@aut@@ Yuan, Kun @@aut@@ Wang, Shang @@aut@@ |
publishDateDaySort_date |
2019-01-03T00:00:00Z |
hierarchy_top_id |
165669608 |
dewey-sort |
14 |
id |
OLC2025619855 |
language_de |
englisch |
fullrecord |
<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim"><record><leader>01000caa a22002652 4500</leader><controlfield tag="001">OLC2025619855</controlfield><controlfield tag="003">DE-627</controlfield><controlfield tag="005">20230504132905.0</controlfield><controlfield tag="007">tu</controlfield><controlfield tag="008">200820s2019 xx ||||| 00| ||eng c</controlfield><datafield tag="024" ind1="7" ind2=" "><subfield code="a">10.1007/s00521-018-3954-7</subfield><subfield code="2">doi</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-627)OLC2025619855</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-He213)s00521-018-3954-7-p</subfield></datafield><datafield tag="040" ind1=" " ind2=" "><subfield code="a">DE-627</subfield><subfield code="b">ger</subfield><subfield code="c">DE-627</subfield><subfield code="e">rakwb</subfield></datafield><datafield tag="041" ind1=" " ind2=" "><subfield code="a">eng</subfield></datafield><datafield tag="082" ind1="0" ind2="4"><subfield code="a">004</subfield><subfield code="q">VZ</subfield></datafield><datafield tag="100" ind1="1" ind2=" "><subfield code="a">Zhu, Liping</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="245" ind1="1" ind2="0"><subfield code="a">Crowd density estimation based on classification activation map and patch density level</subfield></datafield><datafield tag="264" ind1=" " ind2="1"><subfield code="c">2019</subfield></datafield><datafield tag="336" ind1=" " ind2=" "><subfield code="a">Text</subfield><subfield code="b">txt</subfield><subfield code="2">rdacontent</subfield></datafield><datafield tag="337" ind1=" " ind2=" "><subfield code="a">ohne Hilfsmittel zu benutzen</subfield><subfield code="b">n</subfield><subfield code="2">rdamedia</subfield></datafield><datafield tag="338" ind1=" " ind2=" "><subfield code="a">Band</subfield><subfield code="b">nc</subfield><subfield code="2">rdacarrier</subfield></datafield><datafield tag="500" ind1=" " ind2=" "><subfield code="a">© Springer-Verlag London Ltd., part of Springer Nature 2019</subfield></datafield><datafield tag="520" ind1=" " ind2=" "><subfield code="a">Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods.</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Crowd density estimation</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Image patch</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Density level</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Attention mechanism</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Classification activation map</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Li, Chengyang</subfield><subfield code="0">(orcid)0000-0002-1379-1222</subfield><subfield code="4">aut</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Yang, Zhongguo</subfield><subfield code="4">aut</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Yuan, Kun</subfield><subfield code="4">aut</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Wang, Shang</subfield><subfield code="4">aut</subfield></datafield><datafield tag="773" ind1="0" ind2="8"><subfield code="i">Enthalten in</subfield><subfield code="t">Neural computing & applications</subfield><subfield code="d">Springer London, 1993</subfield><subfield code="g">32(2019), 9 vom: 03. Jan., Seite 5105-5116</subfield><subfield code="w">(DE-627)165669608</subfield><subfield code="w">(DE-600)1136944-9</subfield><subfield code="w">(DE-576)032873050</subfield><subfield code="x">0941-0643</subfield><subfield code="7">nnns</subfield></datafield><datafield tag="773" ind1="1" ind2="8"><subfield code="g">volume:32</subfield><subfield code="g">year:2019</subfield><subfield code="g">number:9</subfield><subfield code="g">day:03</subfield><subfield code="g">month:01</subfield><subfield code="g">pages:5105-5116</subfield></datafield><datafield tag="856" ind1="4" ind2="1"><subfield code="u">https://doi.org/10.1007/s00521-018-3954-7</subfield><subfield code="z">lizenzpflichtig</subfield><subfield code="3">Volltext</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_USEFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SYSFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_OLC</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SSG-OLC-MAT</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_70</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_2018</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4277</subfield></datafield><datafield tag="951" ind1=" " ind2=" "><subfield code="a">AR</subfield></datafield><datafield tag="952" ind1=" " ind2=" "><subfield code="d">32</subfield><subfield code="j">2019</subfield><subfield code="e">9</subfield><subfield code="b">03</subfield><subfield code="c">01</subfield><subfield code="h">5105-5116</subfield></datafield></record></collection>
|
author |
Zhu, Liping |
spellingShingle |
Zhu, Liping ddc 004 misc Crowd density estimation misc Image patch misc Density level misc Attention mechanism misc Classification activation map Crowd density estimation based on classification activation map and patch density level |
authorStr |
Zhu, Liping |
ppnlink_with_tag_str_mv |
@@773@@(DE-627)165669608 |
format |
Article |
dewey-ones |
004 - Data processing & computer science |
delete_txt_mv |
keep |
author_role |
aut aut aut aut aut |
collection |
OLC |
remote_str |
false |
illustrated |
Not Illustrated |
issn |
0941-0643 |
topic_title |
004 VZ Crowd density estimation based on classification activation map and patch density level Crowd density estimation Image patch Density level Attention mechanism Classification activation map |
topic |
ddc 004 misc Crowd density estimation misc Image patch misc Density level misc Attention mechanism misc Classification activation map |
topic_unstemmed |
ddc 004 misc Crowd density estimation misc Image patch misc Density level misc Attention mechanism misc Classification activation map |
topic_browse |
ddc 004 misc Crowd density estimation misc Image patch misc Density level misc Attention mechanism misc Classification activation map |
format_facet |
Aufsätze Gedruckte Aufsätze |
format_main_str_mv |
Text Zeitschrift/Artikel |
carriertype_str_mv |
nc |
hierarchy_parent_title |
Neural computing & applications |
hierarchy_parent_id |
165669608 |
dewey-tens |
000 - Computer science, knowledge & systems |
hierarchy_top_title |
Neural computing & applications |
isfreeaccess_txt |
false |
familylinks_str_mv |
(DE-627)165669608 (DE-600)1136944-9 (DE-576)032873050 |
title |
Crowd density estimation based on classification activation map and patch density level |
ctrlnum |
(DE-627)OLC2025619855 (DE-He213)s00521-018-3954-7-p |
title_full |
Crowd density estimation based on classification activation map and patch density level |
author_sort |
Zhu, Liping |
journal |
Neural computing & applications |
journalStr |
Neural computing & applications |
lang_code |
eng |
isOA_bool |
false |
dewey-hundreds |
000 - Computer science, information & general works |
recordtype |
marc |
publishDateSort |
2019 |
contenttype_str_mv |
txt |
container_start_page |
5105 |
author_browse |
Zhu, Liping Li, Chengyang Yang, Zhongguo Yuan, Kun Wang, Shang |
container_volume |
32 |
class |
004 VZ |
format_se |
Aufsätze |
author-letter |
Zhu, Liping |
doi_str_mv |
10.1007/s00521-018-3954-7 |
normlink |
(ORCID)0000-0002-1379-1222 |
normlink_prefix_str_mv |
(orcid)0000-0002-1379-1222 |
dewey-full |
004 |
title_sort |
crowd density estimation based on classification activation map and patch density level |
title_auth |
Crowd density estimation based on classification activation map and patch density level |
abstract |
Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods. © Springer-Verlag London Ltd., part of Springer Nature 2019 |
abstractGer |
Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods. © Springer-Verlag London Ltd., part of Springer Nature 2019 |
abstract_unstemmed |
Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods. © Springer-Verlag London Ltd., part of Springer Nature 2019 |
collection_details |
GBV_USEFLAG_A SYSFLAG_A GBV_OLC SSG-OLC-MAT GBV_ILN_70 GBV_ILN_2018 GBV_ILN_4277 |
container_issue |
9 |
title_short |
Crowd density estimation based on classification activation map and patch density level |
url |
https://doi.org/10.1007/s00521-018-3954-7 |
remote_bool |
false |
author2 |
Li, Chengyang Yang, Zhongguo Yuan, Kun Wang, Shang |
author2Str |
Li, Chengyang Yang, Zhongguo Yuan, Kun Wang, Shang |
ppnlink |
165669608 |
mediatype_str_mv |
n |
isOA_txt |
false |
hochschulschrift_bool |
false |
doi_str |
10.1007/s00521-018-3954-7 |
up_date |
2024-07-04T01:43:49.459Z |
_version_ |
1803610939152400384 |
fullrecord_marcxml |
<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim"><record><leader>01000caa a22002652 4500</leader><controlfield tag="001">OLC2025619855</controlfield><controlfield tag="003">DE-627</controlfield><controlfield tag="005">20230504132905.0</controlfield><controlfield tag="007">tu</controlfield><controlfield tag="008">200820s2019 xx ||||| 00| ||eng c</controlfield><datafield tag="024" ind1="7" ind2=" "><subfield code="a">10.1007/s00521-018-3954-7</subfield><subfield code="2">doi</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-627)OLC2025619855</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-He213)s00521-018-3954-7-p</subfield></datafield><datafield tag="040" ind1=" " ind2=" "><subfield code="a">DE-627</subfield><subfield code="b">ger</subfield><subfield code="c">DE-627</subfield><subfield code="e">rakwb</subfield></datafield><datafield tag="041" ind1=" " ind2=" "><subfield code="a">eng</subfield></datafield><datafield tag="082" ind1="0" ind2="4"><subfield code="a">004</subfield><subfield code="q">VZ</subfield></datafield><datafield tag="100" ind1="1" ind2=" "><subfield code="a">Zhu, Liping</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="245" ind1="1" ind2="0"><subfield code="a">Crowd density estimation based on classification activation map and patch density level</subfield></datafield><datafield tag="264" ind1=" " ind2="1"><subfield code="c">2019</subfield></datafield><datafield tag="336" ind1=" " ind2=" "><subfield code="a">Text</subfield><subfield code="b">txt</subfield><subfield code="2">rdacontent</subfield></datafield><datafield tag="337" ind1=" " ind2=" "><subfield code="a">ohne Hilfsmittel zu benutzen</subfield><subfield code="b">n</subfield><subfield code="2">rdamedia</subfield></datafield><datafield tag="338" ind1=" " ind2=" "><subfield code="a">Band</subfield><subfield code="b">nc</subfield><subfield code="2">rdacarrier</subfield></datafield><datafield tag="500" ind1=" " ind2=" "><subfield code="a">© Springer-Verlag London Ltd., part of Springer Nature 2019</subfield></datafield><datafield tag="520" ind1=" " ind2=" "><subfield code="a">Abstract The task of crowd counting and density map estimation is riddled with many challenges, such as occlusions, non-uniform density, intra-scene and inter-scene variations in scale and perspective. Due to the development of deep learning and large crowd datasets in recent years, most crowd counting methods have achieved notable success. This paper aims to solve crowd density estimation problem for both sparse and dense conditions. To this end, we make two contributions: (1) a network named Patch Scale Discriminant Regression Network (PSDR). Given an input crowd image, it divides the image into patches and sends image patches of different density levels into different regression networks to get the corresponding density maps. It combines all patch density maps to predict the entire density map as the output. (2) A person classification activation map (CAM) method. CAM provides person location information and guides the generation of the entire density map in the final stage. Experiment confirms that CAM allows PSDR to gain another round of performance boost. For instance, on the SmartCity dataset, we achieve (8.6–1.1) MAE and (11.6–1.4) MSE. Our method combining above two methods performs better than state-of-the-art methods.</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Crowd density estimation</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Image patch</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Density level</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Attention mechanism</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">Classification activation map</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Li, Chengyang</subfield><subfield code="0">(orcid)0000-0002-1379-1222</subfield><subfield code="4">aut</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Yang, Zhongguo</subfield><subfield code="4">aut</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Yuan, Kun</subfield><subfield code="4">aut</subfield></datafield><datafield tag="700" ind1="1" ind2=" "><subfield code="a">Wang, Shang</subfield><subfield code="4">aut</subfield></datafield><datafield tag="773" ind1="0" ind2="8"><subfield code="i">Enthalten in</subfield><subfield code="t">Neural computing & applications</subfield><subfield code="d">Springer London, 1993</subfield><subfield code="g">32(2019), 9 vom: 03. Jan., Seite 5105-5116</subfield><subfield code="w">(DE-627)165669608</subfield><subfield code="w">(DE-600)1136944-9</subfield><subfield code="w">(DE-576)032873050</subfield><subfield code="x">0941-0643</subfield><subfield code="7">nnns</subfield></datafield><datafield tag="773" ind1="1" ind2="8"><subfield code="g">volume:32</subfield><subfield code="g">year:2019</subfield><subfield code="g">number:9</subfield><subfield code="g">day:03</subfield><subfield code="g">month:01</subfield><subfield code="g">pages:5105-5116</subfield></datafield><datafield tag="856" ind1="4" ind2="1"><subfield code="u">https://doi.org/10.1007/s00521-018-3954-7</subfield><subfield code="z">lizenzpflichtig</subfield><subfield code="3">Volltext</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_USEFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SYSFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_OLC</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SSG-OLC-MAT</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_70</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_2018</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4277</subfield></datafield><datafield tag="951" ind1=" " ind2=" "><subfield code="a">AR</subfield></datafield><datafield tag="952" ind1=" " ind2=" "><subfield code="d">32</subfield><subfield code="j">2019</subfield><subfield code="e">9</subfield><subfield code="b">03</subfield><subfield code="c">01</subfield><subfield code="h">5105-5116</subfield></datafield></record></collection>
|
score |
7.4005365 |