《電子技術(shù)應(yīng)用》
您所在的位置:首頁(yè) > 人工智能 > 設(shè)計(jì)應(yīng)用 > 基于邊緣計(jì)算中極端姿態(tài)和表情的人臉識(shí)別
基于邊緣計(jì)算中極端姿態(tài)和表情的人臉識(shí)別
2021年電子技術(shù)應(yīng)用第6期
況朝青1,2,3,賀 超1,2,3,王均成1,2,3,鄒建紋1,2,3
1.重慶郵電大學(xué) 通信與信息工程學(xué)院,重慶 400065;2.重慶高校市級(jí)光通信與網(wǎng)絡(luò)重點(diǎn)實(shí)驗(yàn)室,重慶 400065; 3.泛在感知與互聯(lián)重慶市重點(diǎn)實(shí)驗(yàn)室,重慶 400065
摘要: 隨著信息技術(shù)的發(fā)展,人臉識(shí)別在支付、工作和安防系統(tǒng)中應(yīng)用的越來(lái)越多。在邊緣計(jì)算系統(tǒng)中,為了處理的速度,通常選擇較小的神經(jīng)網(wǎng)絡(luò)進(jìn)行人臉識(shí)別,這樣會(huì)導(dǎo)致識(shí)別率低。并且在實(shí)際應(yīng)用中大多都是對(duì)于圖片質(zhì)量較高的人臉可以很好地識(shí)別,但對(duì)于受光照影響較大、表情和姿態(tài)變化大的圖片識(shí)別率不是很高。因此,選擇SqueezeNet輕量級(jí)網(wǎng)絡(luò),該網(wǎng)絡(luò)層數(shù)小,可以很好地運(yùn)用于邊緣計(jì)算系統(tǒng)中。采用了預(yù)處理的方法來(lái)對(duì)圖片進(jìn)行預(yù)處理,然后改進(jìn)了SqueezeNet網(wǎng)絡(luò)的損失函數(shù)以及加入了ResNet網(wǎng)絡(luò)中的殘差學(xué)習(xí)方法。最后通過(guò)對(duì)LFW和IJB-A數(shù)據(jù)集進(jìn)行測(cè)試,該研究方法明顯提高了識(shí)別率。
中圖分類號(hào): TN911.73;TP391.4
文獻(xiàn)標(biāo)識(shí)碼: A
DOI:10.16157/j.issn.0258-7998.200968
中文引用格式: 況朝青,賀超,王均成,等. 基于邊緣計(jì)算中極端姿態(tài)和表情的人臉識(shí)別[J].電子技術(shù)應(yīng)用,2021,47(6):30-34.
英文引用格式: Kuang Chaoqing,He Chao,Wang Juncheng,et al. Face recognition with extreme posture and expression[J]. Application of Electronic Technique,2021,47(6):30-34.
Face recognition with extreme posture and expression
Kuang Chaoqing1,2,3,He Chao1,2,3,Wang Juncheng1,2,3,Zou Jianwen1,2,3
1.School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications, Chongqing 400065,China; 2.Optical Communications and Networks Key Laboratory of Chongqing,Chongqing 400065,China; 3.Ubiquitous Sensing and Networking Key Laboratory of Chongqing,Chongqing 400065,China
Abstract: With the development of information technology, face recognition is used more and more in payment, work and security system. In the edge computing system, in order to deal with the speed, we usually choose a smaller neural network for face recognition, which may cause the recognition rate is not very high. And in practical applications, most of them can recognize the face with high image quality, but the recognition rate is not very high for the face which is greatly affected by the light and has great changes in expression and posture. Therefore, this paper chooses the SqueezeNet lightweight network, which has a small number of layers and can be well used in edge computing system. The method of preprocessing is used to preprocess the image, and then the loss function of SqueezeNet network and the residual learning method of ResNet network are improved. Finally, through the test of LFW and IJB-A data set, it is concluded that the research method in this paper can significantly improve the recognition rate.
Key words : neural network;face recognition;preprocessing;SqueezeNet network;ResNet network

0 引言

    近年來(lái),人臉識(shí)別受到越來(lái)越多的關(guān)注,主要是通過(guò)神經(jīng)網(wǎng)絡(luò)模型來(lái)進(jìn)行人臉識(shí)別。但人臉識(shí)別依然是一個(gè)非常重要但又極具挑戰(zhàn)性的問(wèn)題,主要是現(xiàn)在大部分的人臉識(shí)別采用的圖像都是靜態(tài)和質(zhì)量較高的圖片,所以識(shí)別效果很好。但在實(shí)際應(yīng)用中,人臉圖像受到光照、表情和較大的姿態(tài)變化的影響,可能導(dǎo)致識(shí)別率急劇下降。因此,采用一種預(yù)處理的方式來(lái)處理圖片,提高圖片的質(zhì)量,成為了當(dāng)下研究的關(guān)鍵[1]。并且在邊緣計(jì)算系統(tǒng)中,采用大型網(wǎng)絡(luò)來(lái)進(jìn)行人臉識(shí)別是不現(xiàn)實(shí)的,主要是受到處理器的速度和功耗的影響,因此這方面的應(yīng)用成為了研究的熱點(diǎn)。




本文詳細(xì)內(nèi)容請(qǐng)下載:http://ihrv.cn/resource/share/2000003569




作者信息:

況朝青1,2,3,賀  超1,2,3,王均成1,2,3,鄒建紋1,2,3

(1.重慶郵電大學(xué) 通信與信息工程學(xué)院,重慶 400065;2.重慶高校市級(jí)光通信與網(wǎng)絡(luò)重點(diǎn)實(shí)驗(yàn)室,重慶 400065;

3.泛在感知與互聯(lián)重慶市重點(diǎn)實(shí)驗(yàn)室,重慶 400065)




wd.jpg

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