《電子技術(shù)應(yīng)用》
您所在的位置:首頁(yè) > 其他 > 設(shè)計(jì)應(yīng)用 > 基于上下文特征重聚合網(wǎng)絡(luò)的人群計(jì)數(shù)
基于上下文特征重聚合網(wǎng)絡(luò)的人群計(jì)數(shù)
信息技術(shù)與網(wǎng)絡(luò)安全
郝曉亮,楊倩倩,夏殷鋒,彭思凡,殷保群
(中國(guó)科學(xué)技術(shù)大學(xué) 信息科學(xué)技術(shù)學(xué)院,安徽 合肥230027)
摘要: 針對(duì)計(jì)數(shù)問題中人群目標(biāo)尺度的變化問題,提出了一種基于上下文特征重聚合的計(jì)數(shù)算法。將高層網(wǎng)絡(luò)提取的語(yǔ)義信息與底層網(wǎng)絡(luò)提取的人群尺度細(xì)節(jié)信息相結(jié)合,旨在利用淺層網(wǎng)絡(luò)中提取的信息向深層網(wǎng)絡(luò)提取的特征中融入不同尺度的行人目標(biāo)特征,從而融合多種尺度的人群特征回歸出高質(zhì)量的人群密度圖。此外,在ShanghaiTech、UCF_CC_50以及 UCF_QNRF 三個(gè)數(shù)據(jù)集進(jìn)行算法的性能驗(yàn)證,并通過結(jié)構(gòu)實(shí)驗(yàn)驗(yàn)證本文結(jié)構(gòu)的有效性。
中圖分類號(hào): TP391
文獻(xiàn)標(biāo)識(shí)碼: A
DOI: 10.19358/j.issn.2096-5133.2021.07.010
引用格式: 郝曉亮,楊倩倩,夏殷鋒,等. 基于上下文特征重聚合網(wǎng)絡(luò)的人群計(jì)數(shù)[J].信息技術(shù)與網(wǎng)絡(luò)安全,2021,40(7):59-65.
Context-aware feature reaggregation network for crowd counting
Hao Xiaoliang,Yang Qianqian,Xia Yinfeng,Peng Sifan,Yin Baoqun
(School of Information Science and Technology,University of Science and Technology of China,Hefei 230027,China)
Abstract: Aiming at the problem of the target scale-aware in crowd counting, this paper proposes a counting algorithm based on the reaggregation of contextual features. The semantic information extracted by the high-level network is combined with the crowd-scale detail information extracted by the low-level network, and the purpose is to use the information extracted from the low network to integrate the features of different scales into the features extracted from the deep network, thereby fusing multiple scales crowd feature return to a high-quality crowd density map. In addition, this paper performs algorithm performance verification on ShanghaiTech, UCF_CC_50 and UCF_QNRF datasets and the effectiveness of structures is verified through structural experiments.
Key words : crowd counting;context-aware feature enhance;multi-scale feature fusion;density map

0 引言

 在人群計(jì)數(shù)所面臨的諸多難題中,人群尺度變化導(dǎo)致的計(jì)數(shù)性能下降問題備受關(guān)注。圖1所示在人群密度較大的場(chǎng)景中,圖片中不同區(qū)域的人群在分布上存在人頭尺度上的不均衡,對(duì)計(jì)數(shù)準(zhǔn)確性造成嚴(yán)重的影響。為了解決此類問題,本文提出了基于上下文的特征增強(qiáng)方法,提取不同尺度的人頭特征,融合經(jīng)過強(qiáng)化的特征,生成反映不同人頭尺度的密度圖



本文詳細(xì)內(nèi)容請(qǐng)下載:http://ihrv.cn/resource/share/2000003679




作者信息:

郝曉亮,楊倩倩,夏殷鋒,彭思凡,殷保群

(中國(guó)科學(xué)技術(shù)大學(xué) 信息科學(xué)技術(shù)學(xué)院,安徽 合肥230027)


此內(nèi)容為AET網(wǎng)站原創(chuàng),未經(jīng)授權(quán)禁止轉(zhuǎn)載。