Improving MLP-Based Weakly Supervised Crowd-Counting Network via Scale Reasoning and Ranking
MLP-based weakly supervised crowd counting approaches have made significant advancements over the past few years. However, owing to the limited datasets, the current MLP-based methods do not consider the problem of region-to-region dependency in the image. For this, we propose a weakly supervised me...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2024-01-01
|
Series: | Electronics |
Subjects: | |
Online Access: | https://www.mdpi.com/2079-9292/13/3/471 |
_version_ | 1797318873140166656 |
---|---|
author | Ming Gao Mingfang Deng Huailin Zhao Yangjian Chen Yongqi Chen |
author_facet | Ming Gao Mingfang Deng Huailin Zhao Yangjian Chen Yongqi Chen |
author_sort | Ming Gao |
collection | DOAJ |
description | MLP-based weakly supervised crowd counting approaches have made significant advancements over the past few years. However, owing to the limited datasets, the current MLP-based methods do not consider the problem of region-to-region dependency in the image. For this, we propose a weakly supervised method termed SR2. SR2 consists of three parts: scale-reasoning module, scale-ranking module, and regression branch. In particular, the scale-reasoning module extracts and fuses the region-to-region dependency in the image and multiple scale feature, then sends the fused features to the regression branch to obtain estimated counts; the scale-ranking module is used to understand the internal information of the image better and expand the datasets efficiently, which will help to improve the accuracy of the estimated counts in the regression branch. We conducted extensive experiments on four benchmark datasets. The final results showed that our approach has better and higher competing counting performance with respect to other weakly supervised counting networks and with respect to some popular fully supervised counting networks. |
first_indexed | 2024-03-08T03:58:43Z |
format | Article |
id | doaj.art-73aafad72e15456083fc8b101a245ece |
institution | Directory Open Access Journal |
issn | 2079-9292 |
language | English |
last_indexed | 2024-03-08T03:58:43Z |
publishDate | 2024-01-01 |
publisher | MDPI AG |
record_format | Article |
series | Electronics |
spelling | doaj.art-73aafad72e15456083fc8b101a245ece2024-02-09T15:10:20ZengMDPI AGElectronics2079-92922024-01-0113347110.3390/electronics13030471Improving MLP-Based Weakly Supervised Crowd-Counting Network via Scale Reasoning and RankingMing Gao0Mingfang Deng1Huailin Zhao2Yangjian Chen3Yongqi Chen4School of Electrical and Electronic Engineering, Shanghai Institute of Technology, Shanghai 201400, ChinaSchool of Electrical and Electronic Engineering, Shanghai Institute of Technology, Shanghai 201400, ChinaSchool of Electrical and Electronic Engineering, Shanghai Institute of Technology, Shanghai 201400, ChinaSchool of Electrical and Electronic Engineering, Shanghai Institute of Technology, Shanghai 201400, ChinaSchool of Electrical and Electronic Engineering, Shanghai Institute of Technology, Shanghai 201400, ChinaMLP-based weakly supervised crowd counting approaches have made significant advancements over the past few years. However, owing to the limited datasets, the current MLP-based methods do not consider the problem of region-to-region dependency in the image. For this, we propose a weakly supervised method termed SR2. SR2 consists of three parts: scale-reasoning module, scale-ranking module, and regression branch. In particular, the scale-reasoning module extracts and fuses the region-to-region dependency in the image and multiple scale feature, then sends the fused features to the regression branch to obtain estimated counts; the scale-ranking module is used to understand the internal information of the image better and expand the datasets efficiently, which will help to improve the accuracy of the estimated counts in the regression branch. We conducted extensive experiments on four benchmark datasets. The final results showed that our approach has better and higher competing counting performance with respect to other weakly supervised counting networks and with respect to some popular fully supervised counting networks.https://www.mdpi.com/2079-9292/13/3/471weakly supervised countingMLPgraph neural networksranking mechanism |
spellingShingle | Ming Gao Mingfang Deng Huailin Zhao Yangjian Chen Yongqi Chen Improving MLP-Based Weakly Supervised Crowd-Counting Network via Scale Reasoning and Ranking Electronics weakly supervised counting MLP graph neural networks ranking mechanism |
title | Improving MLP-Based Weakly Supervised Crowd-Counting Network via Scale Reasoning and Ranking |
title_full | Improving MLP-Based Weakly Supervised Crowd-Counting Network via Scale Reasoning and Ranking |
title_fullStr | Improving MLP-Based Weakly Supervised Crowd-Counting Network via Scale Reasoning and Ranking |
title_full_unstemmed | Improving MLP-Based Weakly Supervised Crowd-Counting Network via Scale Reasoning and Ranking |
title_short | Improving MLP-Based Weakly Supervised Crowd-Counting Network via Scale Reasoning and Ranking |
title_sort | improving mlp based weakly supervised crowd counting network via scale reasoning and ranking |
topic | weakly supervised counting MLP graph neural networks ranking mechanism |
url | https://www.mdpi.com/2079-9292/13/3/471 |
work_keys_str_mv | AT minggao improvingmlpbasedweaklysupervisedcrowdcountingnetworkviascalereasoningandranking AT mingfangdeng improvingmlpbasedweaklysupervisedcrowdcountingnetworkviascalereasoningandranking AT huailinzhao improvingmlpbasedweaklysupervisedcrowdcountingnetworkviascalereasoningandranking AT yangjianchen improvingmlpbasedweaklysupervisedcrowdcountingnetworkviascalereasoningandranking AT yongqichen improvingmlpbasedweaklysupervisedcrowdcountingnetworkviascalereasoningandranking |