基于改進(jìn)AlexNet卷積神經(jīng)網(wǎng)絡(luò)人臉識別的研究
電子技術(shù)應(yīng)用
蔡靖,谷承睿,劉光達(dá),孫慧慧
吉林大學(xué) 儀器科學(xué)與電氣工程學(xué)院
摘要: 近期,人臉識別技術(shù)在社會上廣受關(guān)注,其非接觸式的識別特性相較于指紋等傳統(tǒng)接觸式識別方法展現(xiàn)出明顯優(yōu)勢。在深度學(xué)習(xí)領(lǐng)域,由于傳統(tǒng)卷積神經(jīng)網(wǎng)絡(luò)在人臉識別任務(wù)上的準(zhǔn)確性和速度尚有提升空間,因此提出采用改進(jìn)的AlexNet卷積神經(jīng)網(wǎng)絡(luò)進(jìn)行人臉識別。通過實(shí)驗驗證,與傳統(tǒng)卷積網(wǎng)絡(luò)相比,改進(jìn)后的AlexNet在人臉識別上不僅準(zhǔn)確度更高,而且識別過程更為穩(wěn)定。
中圖分類號:TN402 文獻(xiàn)標(biāo)志碼:A DOI: 10.16157/j.issn.0258-7998.245086
中文引用格式: 蔡靖,谷承睿,劉光達(dá),等. 基于改進(jìn)AlexNet卷積神經(jīng)網(wǎng)絡(luò)人臉識別的研究[J]. 電子技術(shù)應(yīng)用,2024,50(11):42-46.
英文引用格式: Cai Jing,Gu Chengrui,Liu Guangda,et al. Research on face recognition based on improved AlexNet convolutional neural network[J]. Application of Electronic Technique,2024,50(11):42-46.
中文引用格式: 蔡靖,谷承睿,劉光達(dá),等. 基于改進(jìn)AlexNet卷積神經(jīng)網(wǎng)絡(luò)人臉識別的研究[J]. 電子技術(shù)應(yīng)用,2024,50(11):42-46.
英文引用格式: Cai Jing,Gu Chengrui,Liu Guangda,et al. Research on face recognition based on improved AlexNet convolutional neural network[J]. Application of Electronic Technique,2024,50(11):42-46.
Research on face recognition based on improved AlexNet convolutional neural network
Cai Jing,Gu Chengrui,Liu Guangda,Sun Huihui
College of Instrumentation & Electrical Engineering, Jilin University
Abstract: In recent years, face recognition has been a hot topic in society. Compared to contact-based recognition methods such as fingerprint recognition, face recognition offers the advantage of being contactless. In the field of deep learning, traditional convolutional neural networks do not achieve high enough accuracy or speed for face recognition. Therefore, this paper proposes a face recognition algorithm using the AlexNet convolutional neural network. Experimental results show that AlexNet provides higher accuracy and more stability in face recognition compared to traditional convolutional neural networks.
Key words : deep learning;convolutional neural network;face recognition;AlexNet
引言
在過去的十年中,卷積神經(jīng)網(wǎng)絡(luò)(CNN)在視覺識別的任務(wù)中取得了很多的成果。人臉識別作為計算機(jī)視覺領(lǐng)域的一個熱點(diǎn)問題,一直是諸多安全和個人識別應(yīng)用的核心技術(shù)[1]。近年來,隨著大數(shù)據(jù)和高性能計算資源的普及,基于深度學(xué)習(xí)的人臉識別技術(shù)已經(jīng)取得了顯著的進(jìn)展。在這些方法中,基于AlexNet架構(gòu)的CNN模型因其強(qiáng)大的特征學(xué)習(xí)能力而備受關(guān)注。本文旨在探討AlexNet卷積神經(jīng)網(wǎng)絡(luò)在人臉識別任務(wù)上的應(yīng)用。通過實(shí)驗驗證了AlexNet在ORL人臉數(shù)據(jù)庫上的性能,并且進(jìn)一步討論了通過遷移學(xué)習(xí)對AlexNet進(jìn)行微調(diào)來適應(yīng)人臉識別任務(wù)的方法,并探索了數(shù)據(jù)增強(qiáng)、網(wǎng)絡(luò)結(jié)構(gòu)調(diào)整和超參數(shù)優(yōu)化對模型性能的影響。在傳統(tǒng)的AlexNet卷積網(wǎng)絡(luò)上進(jìn)行改進(jìn),得到優(yōu)化后的結(jié)果。
本文詳細(xì)內(nèi)容請下載:
http://ihrv.cn/resource/share/2000006208
作者信息:
蔡靖,谷承睿,劉光達(dá),孫慧慧
(吉林大學(xué) 儀器科學(xué)與電氣工程學(xué)院,吉林 長春 130022)
此內(nèi)容為AET網(wǎng)站原創(chuàng),未經(jīng)授權(quán)禁止轉(zhuǎn)載。