基于IASPP-ResNet分割算法的手势识别

doi:10.12263/DZXB.20210049

电子学报

• •

基于IASPP-ResNet分割算法的手势识别

雷玉¹^,², 崔振超¹^,²(), 陈丽萍¹, 陈向阳¹, 王煜骁¹

^1.河北大学网络空间安全与计算机学院，河北保定 071002
^2.河北省机器视觉工程研究中心，河北保定 071002

收稿日期:2020-12-31 修回日期:2021-03-27 出版日期:2022-07-04
通讯作者: 崔振超
作者简介:雷玉女，1995年生，山西晋中人.河北大学网络空间安全与计算机学院硕士研究生，研究方向为深度学习、图像分割技术及其应用.
崔振超（通讯作者）男，1983年生，河北邯郸人，讲师.中国计算机学会会员，2007年于燕山大学获得学士学位，2010年于燕山大学获得硕士学位，2015年于哈尔滨工业大学获得博士学位.现为河北大学网络空间安全与计算机学院教师，主要从事人工智能、机器视觉方向研究. E-mail: cuizhenchao@gmail.com
陈丽萍女，1974年生，河北保定人，讲师.1997年于河北农业大学获得学士学位，2000年获得硕士学位.现为河北大学网络空间安全与计算机学院教师，主要从事机器视觉方面的研究.
陈向阳女，1977年生，河南三门峡人，讲师.2000年毕业于燕山大学获得学士学位，2007年毕业于河北大学获得硕士学位.现为河北大学网络空间安全与计算机学院教师，研究方向为深度学习.
王煜骁男，1997年生，河北廊坊人.河北大学网络空间安全与计算机学院硕士研究生，研究方向为深度学习、图像分类.
基金资助:
河北省自然科学基金项目(F2017201069);河北省研究生创新资助项目(HBU2021ss059)

Hand Gesture Recognition Based on IASPP-ResNet Segmentation Algorithm

LEI Yu¹^,², CUI Zhen-chao¹^,²(), CHEN Li-ping¹, CHEN Xiang-yang¹, WANG Yu-xiao¹

^1.School of Cyber Security and Computer，Hebei University，Baoding，Hebei 071002，China
^2.Hebei Machine Vision Engineering Research Center，Baoding，Hebei 071002，China

Received:2020-12-31 Revised:2021-03-27 Online:2022-07-04
Contact: CUI Zhen-chao

摘要/Abstract

摘要：

手势识别是计算机视觉领域中研究的重要领域，是人机交互领域的重要组成.由于其识别结果受到复杂背景的影响，手势识别面临着巨大挑战.为了解决复杂背景影响的问题，本文利用了密集分割+手势分类的组合型模型，提出了一种新的手势识别算法.在密集分割部分，本文提出了改进型空洞空间金字塔池化（Improved Atrous Spatial Pyramid Pooling， IASPP）.IASPP通过密集的连接不同空洞率的空洞卷积获取了在不同视野上的手势多尺度信息，从而提高了特征表述的精确性.另外，为了融合不同层级上的细节和空间位置信息，提升整体网络的分割性能，本文将IASPP嵌入编码器-解码器结构中，提出了IASPP-ResNet手势分割算法.在手势识别部分，我们利用了深度卷积神经网络模型，获得了较高的识别率.实验结果表明，在目前常用的公共数据集上，与传统的机器学习方法以及基于深度学习的方法相比，IASPP-ResNet分割算法的准确率更高，并且本文提出的密集分割+手势分类的组合型模型在NUS-II数据集上的手势识别率可达98.63%，优于现有的手势识别算法.

关键词: 手势分割, 复杂背景, 编码器-解码器, 空洞卷积, 手势识别

Abstract:

Gesture recognition is an essential research area in the field of computer vision, and it is also a significant component of the human-computer interaction. Due to its recognition results can be influenced by complex backgrounds, gesture recognition faces huge challenges. To solve the problem that is affected by the complex background, this paper proposes a new gesture recognition algorithm based on the combined model of dense segmentation and gesture classification. In the dense segmentation part, this paper shows the Improved Atrous Spatial Pyramid Pooling (IASPP). IASPP is a pooling layer in a convolution neural network, which can obtain the refine features by connecting cascade model and parallel model in atrous spatial pyramid pooling. Otherwise, in order to improve the segmentation performance by integrating details and spatial location information at different levels, the IASPP was embedded in a ResNet with encoder-decoder, and we name the method the Improved Atrous Spatial Pyramid Pooling-ResNet (IASPP-ResNet) for gesture segmentation. In the part of gesture recognition, we use the deep convolutional neural network model to obtain a higher recognition rate. The experimental results show that the IASPP-ResNet segmentation algorithm has a higher accuracy rate on the commonly used public data sets, compared with the traditional machine learning methods as well as the deep learning-based methods, and the gesture recognition rate of the combined model of dense segmentation and gesture classification proposed in this paper can reach 98.63% on NUS-II dataset, which is superior to the existing gesture recognition algorithm.

Key words: hand gesture segmentation, complex background, encoder-decoder, atrous convolution, gesture recognition

中图分类号:

TP391.4

雷玉, 崔振超, 陈丽萍, 陈向阳, 王煜骁. 基于IASPP-ResNet分割算法的手势识别[J]. 电子学报, DOI: 10.12263/DZXB.20210049.

LEI Yu, CUI Zhen-chao, CHEN Li-ping, CHEN Xiang-yang, WANG Yu-xiao. Hand Gesture Recognition Based on IASPP-ResNet Segmentation Algorithm[J]. Acta Electronica Sinica, DOI: 10.12263/DZXB.20210049.

图/表 18

图1 IASPP框架图:图中的d表示空洞率大小,k表示卷积核大小,Concat 表示特征图在通道上进行拼接.

图2 模型的特征金字塔.(a) ASPP模型;(b) IASPP模型.其中,每个条带中的数字表示不同空洞率的组合,每个条带的长度表示每个组合的等效核大小,条带右边的k表示相应的感受野大小.

图3 空洞卷积堆叠图:(a)d=8的空洞卷积;(b)将d=4的空洞卷积堆叠在d=8的空洞卷积之上;(c)将d=2,4的空洞卷积堆叠在d=8的空洞卷积之上.其中橙色表示信息的来源.

图4 IASPP-ResNet框架

表1 全卷积深度残差网络参数设置

Layer name	Output shape	layers
input	512 $×$ 512 $×$ 3
Convolution1	256 $×$ 256 $×$ 64
Convolution2	128 $×$ 128 $×$ 64
ResGroup1	128 $×$ 128 $×$ 256	$1 × 1 64 3 × 3 64 1 × 1 256 × 3$ ,d=1,s=2
ResGroup2	64 $×$ 64 $×$ 512	$1 × 1 128 3 × 3 128 1 × 1 512 × 4$ ,d=1,s=1
ResGroup3	64 $×$ 64 $×$ 1024	$1 × 1 256 3 × 3 256 1 × 1 1024 × 6$ ,d=2,s=1
ResGroup4	64 $×$ 64 $×$ 2048	$1 × 1 512 3 × 3 512 1 × 1 2048 × 3$ ,d=4,s=1

表1 全卷积深度残差网络参数设置

Layer name	Output shape	layers
input	512 $×$ 512 $×$ 3
Convolution1	256 $×$ 256 $×$ 64
Convolution2	128 $×$ 128 $×$ 64
ResGroup1	128 $×$ 128 $×$ 256	$1 × 1 64 3 × 3 64 1 × 1 256 × 3$ ,d=1,s=2
ResGroup2	64 $×$ 64 $×$ 512	$1 × 1 128 3 × 3 128 1 × 1 512 × 4$ ,d=1,s=1
ResGroup3	64 $×$ 64 $×$ 1024	$1 × 1 256 3 × 3 256 1 × 1 1024 × 6$ ,d=2,s=1
ResGroup4	64 $×$ 64 $×$ 2048	$1 × 1 512 3 × 3 512 1 × 1 2048 × 3$ ,d=4,s=1

图5 整体框架

表2 本文算法与机器学习方法在HGR数据集下的对比结果

Method	Pr	Re	F-score	ROC-area
DSPF^[30]	0.8297	0.9531	0.8552	0.8546
DTF^[31]	0.7816	0.9596	0.8166	0.8823
Bayes^[32]	0.6998	0.9297	0.7422	0.7341
RC^[33]	0.8085	0.8097	0.8087	0.5798
Conaire^[5]	0.7858	0.8228	0.7940	0.7034
Ours	0.9948	0.9929	0.9939	0.9982

图6 本文算法与机器学习方法的测试结果柱状图

表3 在OUTHANDS数据集上的测试结果

Method	MIoU	PA	Pr	Re	F-score	Time
FCN-8s	0.783 0	0.939 3	0.993 9	0.936 7	0.964 4	73.05
PSPNet	0.813 4	0.950 6	0.994 2	0.949 2	0.971 2	92.26
ICNET[34]	0.772 8	0.935 1	0.988 4	0.937 5	0.962 3	39.05
Deeplabv3	0.797 0	0.944 2	0.994 3	0.941 9	0.967 4	113.74
Deeplabv3+	0.887 2	0.973 6	0.990 4	0.979 5	0.984 9	114.66
DUNet[35]	0.892 9	0.974 1	0.990 6	0.979 9	0.985 2	82.59
SPNet[36]	0.893 1	0.976 3	0.991 0	0.981 3	0.986 1	96.23
IASPP	0.896 0	0.977 5	0.990 7	0.982 0	0.986 3	121.01

表4 在HGR数据集上的测试结果

Method	MIoU	PA	Pr	Re	F-score
FCN-8s	0.816 8	0.937 6	0.926 0	0.997 6	0.960 5
PSPNet	0.892 4	0.960 1	0.996 8	0.950 2	0.972 9
ICNET[34]	0.885 4	0.957 3	0.994 8	0.948 7	0.971 2
Deeplabv3	0.882 6	0.956 0	0.997 4	0.944 5	0.970 2
Deeplabv3+	0.959 4	0.985 5	0.993 0	0.987 4	0.990 2
DUNet[35]	0.961 1	0.986 3	0.995 0	0.986 6	0.990 8
SPNet[36]	0.971 2	0.990 1	0.993 4	0.990 1	0.991 7
IASPP	0.975 2	0.990 9	0.994 8	0.992 9	0.993 9

图7 OUTHANDS数据集上各算法测试结果

图8 HGR数据集上各算法测试结果

表5 在OUTHANDS数据集上的识别率

Method	Time In OUTHANDS	Accuracy
文献[25]	/	0.946 2
Deeplabv3+文献[25]	123.65	0.965 3
DUNet+文献[25]	92.60	0.979 4
SPNet+文献[25]	109.26	0.980 9
本文	130.42	0.986 1

表6 在NUS-II数据集上的识别率

Method	Accuracy
文献[26]	0.965 3
文献[37]	0.983 0
DAG-CNN[38]	0.910 5
CNN-SPP[39]	0.959 5
文献[25]	0.966 2
本文	0.986 3

表7 在HGR数据集上的测试结果

Rate	MIoU	F-score
{ $20, 21, 22, 23, 24$ }	0.975 2	0.993 9
{ $20, 21, 22, 21 + 23, 22 + 24$ }	0.975 3	0.993 9
{ $20, 22, 23, 22 + 23, 22 + 24$ }	0.973 4	0.993 5
{ $20, 21, 21 + 22, 21 + 23, 24$ }	0.974 1	0.993 7

表7 在HGR数据集上的测试结果

Rate	MIoU	F-score
{ $20, 21, 22, 23, 24$ }	0.975 2	0.993 9
{ $20, 21, 22, 21 + 23, 22 + 24$ }	0.975 3	0.993 9
{ $20, 22, 23, 22 + 23, 22 + 24$ }	0.973 4	0.993 5
{ $20, 21, 21 + 22, 21 + 23, 24$ }	0.974 1	0.993 7

表8 在OUTHANDS数据集上的测试结果

Rate	MIoU	F-score
{ $20, 21, 22, 23, 24$ }	0.896 0	0.986 3
{ $20, 21, 22, 21 + 23, 22 + 24$ }	0.912 5	0.989 1
{ $20, 22, 23, 22 + 23, 22 + 24$ }	0.899 6	0.987 7
{ $20, 21, 21 + 22, 21 + 23, 24$ }	0.900 2	0.987 2

表8 在OUTHANDS数据集上的测试结果

Rate	MIoU	F-score
{ $20, 21, 22, 23, 24$ }	0.896 0	0.986 3
{ $20, 21, 22, 21 + 23, 22 + 24$ }	0.912 5	0.989 1
{ $20, 22, 23, 22 + 23, 22 + 24$ }	0.899 6	0.987 7
{ $20, 21, 21 + 22, 21 + 23, 24$ }	0.900 2	0.987 2

表9 不同网络架构对OUTHANDS数据集的影响

Rate	MIoU	PA	F-score
DCNN+并行、级联模式+解码器	0.894 8	0.975 9	0.985 3
DCNN+并行模型、反卷积+解码器	0.892 0	0.973 2	0.983 3
DCNN+ IASPP+新的解码器	0.902 9	0.976 4	0.986 6
DCNN+IASPP+解码器(本文模型)	0.896 0	0.977 5	0.986 3

表10 不同网络架构对HGR数据集的影响

Rate	MIoU	PA	F-score
DCNN+并行、级联模式+解码器	0.973 7	0.987 8	0.993 9
DCNN+并行模型、反卷积+解码器	0.971 8	0.984 9	0.993 4
DCNN+ IASPP+新的解码器	0.975 0	0.990 9	0.993 8
DCNN+IASPP+解码器(本文模型)	0.975 2	0.990 9	0.993 9

参考文献 40

1	王勇, 王沙沙, 田增山, 等. 基于FMCW雷达的双流融合神经网络手势识别方法[J].电子学报, 2019, 47(7): 1408-1415.
	WANG Y, WANG S S, TIAN Z S, et al. Two-stream fusion neural network approach for hand gesture recognition based on FMCW radar[J]. Acta Electronica Sinica, 2019, 47(7): 1408-1415. (in Chinese)
2	SAYED U, MOFADDEL M A, BAKHEET S, et al. An elliptical boundary skin model for hand detection based on HSV color space[J]. Information Sciences Letters, 2018, 7(1):13-17.
3	LIU C, WANG J, ZHANG T, et al. Adaptive threshold gesture segmentation algorithm based on skin color[C]//Proceedings of 2016 2nd International Conference on Advances in Mechanical Engineering and Industrial Informatics(AMEII 2016). Hangzhou: Computer Science and Electronic Technology International Society, 2016: 1602-1605.
4	ZHENG Y, ZHENG P. Hand segmentation based on improved gaussian mixture model[C]//2015 International Conference on Computer Science and Applications (CSA). Wuhan: IEEE, 2015:168-171.
5	CONAIRE C O, O'CONNOR N E, SMEATON A F. Detector adaptation by maximising agreement between independent data sources[C]//Conference on Computer Vision and Pattern Recognition. Minneapolis, MN, USA: IEEE, 2007: 1-6.
6	WANG X, FANG Y, LI C, et al. Static gesture segmentation technique based on improved sobel operator[J]. The Journal of Engineering, 2019, 2019(22).
7	TOFIGHI G, MONADJEMI S A, GHASEM-AGHAEE N. Rapid hand posture recognition using adaptive histogram template of skin and hand edge contour[C]//Iranian Conference on Machine Vision and Image Processing. Isfahan: IEEE, 2010: 1-5.
8	CHEN D, LI G, SUN Y, et al. Fusion hand gesture segmentation and extraction based on CMOS sensor and 3D sensor[J]. International Journal of Wireless and Mobile Computing, 2017, 12(3): 305-312.
9	LONG J, SHELHAMER E, DARRELL T. Fully convolutional networks for semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston:IEEE, 2015: 3431-3440.
10	RONNEBERGER O, FISCHER P, BROX T. U-net: Convolutional networks for biomedical image segmentation[C]//International Conference on Medical Image Computing and Computer-assisted Intervention. Cham: Springer International Publishing, 2015: 234-241.
11	ZHAO H, SHI J, QI X, et al. Pyramid scene parsing network[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Honolulu: IEEE, 2017: 2881-2890.
12	CHEN L C, PAPANDREOU G, KOKKINOS I, et al. Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 40(4): 834-848.
13	CHEN L C, PAPANDREOU G, SCHROFF F, et al. Rethinking atrous convolution for semantic image segmentation[EB/OL]. (2017-12-05)[2020-12-31]. .
14	CHEN L C, ZHU Y, PAPANDREOU G, et al. Encoder-decoder with atrous separable convolution for semantic image segmentation[C]//Proceedings of the European Conference on Computer Vision. Munich: Springer, 2018: 801-818.
15	DADASHZADEH A, TARGHI A T, TAHMASBI M, et al. HGR-Net: A fusion network for hand gesture segmentation and recognition[J]. IET Computer Vision, 2019, 13(8): 700-707.
16	卫保国, 徐勇, 刘金玮, 等. 融合SSD目标检测的自适应手势分割方法[J]. 信号处理, 2020, 36(07): 1038-1047.
	WEI B G, XU Y, LIU J W, et al. Adaptive gesture segmentation based on SSD object detection[J]. Journal of Signal Processing, 2020, 36(7): 1038-1047. (in Chinese)
17	MOHANTY A, RAMBHATLA S S, SAHAY R R. Deep gesture: Static hand gesture recognition using CNN[C]//Proceedings of International Conference on Computer Vision and Image Processing. Singapore: Springer, 2017: 449-461.
18	YANG H L, XUAN S B, MO Y B. Hand gesture recognition based on convolution neural network[J]. Cluster Computing, 2019, 22(2): 2719-2729.
19	XING K, DING Z, JIANG S, et al. Hand gesture recognition based on deep learning method[C]//2018 IEEE Third International Conference on Data Science in Cyberspace (DSC). Guangzhou: IEEE, 2018: 542-546.
20	AMEEN S, VADERA S. A convolutional neural network to classify American Sign Language fingerspelling from depth and colour images[J]. Expert Systems, 2017, 34(3): e12197.
21	YU F, KOLTUN V. Multi-scale context aggregation by dilated convolutions[EB/OL]. (2016-04-30)[2020-12-31]. .
22	YANG M K, YU K, ZHANG C, et al. DenseASPP for semantic segmentation in street scenes[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City: IEEE, 2018. 3684-3692.
23	李宝奇, 贺昱曜, 何灵蛟, 等. 基于全卷积神经网络的非对称并行语义分割模型[J]. 电子学报, 2019, 47(5): 1058-1064.
	LI B Q, HE Y Y, HE L J, et al. Asymmetric parallel semantic segmentation model based on full convolutional neural network[J]. Acta Electronica Sinica, 2019, 47(5): 1058-1064. (in Chinese)
24	孟琭, 徐磊, 郭嘉阳. 一种基于改进的MobileNetV2网络语义分割算法[J]. 电子学报, 2020, 48(9): 1769-1776.
	MENG L, XU L, GUO J Y. Semantic segmentation algorithm based on improved MobileNetV2[J]. Acta Electronica Sinica, 2020, 48(9): 1769-1776. (in Chinese)
25	ADITHYA V, RAJESH R. A deep convolutional neural network approach for static hand gesture recognition[J]. Procedia Computer Science, 2020, 171:2353-2361.
26	ZHANG Q, YANG M, KPALMA K, et al. Segmentation of hand posture against complex backgrounds based on saliency and skin colour detection[J]. IAENG International Journal of Computer Science, 2018, 45(3): 435-444.
27	罗会兰, 张云. 基于深度网络的图像语义分割综述[J]. 电子学报, 2019, 47(10): 2211-2220.
	LUO H L, ZHANG Y. A survey of image semantic segmentation based on deep network[J]. Acta Electronica Sinica, 2019, 47(10): 2211-2220.(in Chinese)
28	张庆锐. 复杂场景下的手势分割算法研究[D]. 山东: 山东大学, 2018.
	ZHANG Q R. Research on Hand Gesture Segmentation Algorithm with Complex Background[D]. Shan Dong: Shandong University, 2018. (in Chinese)
29	GARCIA-GARCIA A, ORTS-ESCOLANO S, OPREA S, et al. A review on deep learning techniques applied to semantic segmentation[EB/OL]. (2017-04-22)[2020-12-31]. .
30	KAWULOK M, KAWULOK J, NALEPA J. Spatial-based skin detection using discriminative skin-presence features[J]. Pattern Recognition Letters, 2014, 41: 3-13.
31	KAWULOK M, KAWULOK J, SMOLKA B. Discriminative textural features for image and video colorization[J]. IEICE Transactions on Information and Systems, 2012, 95(7): 1722-1730.
32	JONES M J, REHG J M. Statistical color models with application to skin detection[J]. International Journal of Computer Vision, 2002, 46(1): 81-96.
33	CHENG M M, MITRA N J, HUANG X, et al. Global contrast based salient region detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 37(3): 569-582.
34	ZHAO H, QI X, SHEN X, et al. Icnet for real-time semantic segmentation on high-resolution images[C]//Proceedings of the European Conference on Computer Vision. Munich: Springer, 2018: 405-420.
35	TIAN Z, HE T, SHEN C H, et al. Decoders matter for semantic segmentation: Data-dependent decoding enables flexible feature aggregation[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach: IEEE, 2019: 3126-3135.
36	HOU Q B, ZHANG L, CHENG M M, et al. Strip pooling: Rethinking spatial pooling for scene parsing[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle: IEEE, 2020: 4002-4011
37	SUN J H, JI T T, ZHANG S B,et al. Research on the hand gesture recognition based on deep learning[C]//2018 12th International Symposium on Antennas, Propagation and EM Theory (ISAPE). Hangzhou: IEEE, 2018: 1-4.
38	ARENAS J O P, MORENO R J, BELEÑO R D H. Convolutional neural network with a DAG architecture for control of a robotic arm by means of hand gestures[J]. Contemporary Engineering Sciences, 2018, 11(12): 547-557.
39	TAN Y S, LIM K M, TEE C, et al. Convolutional neural network with spatial pyramid pooling for hand gesture recognition[J]. Neural Computing and Applications, 2020: 1-13.
40	MEHTA S, RASTEGARI M, CASPI A, et al. Espnet: Eficient spatial pyramid of dilated convolutions for semantic segmentation[C]//Proceedings of the European Conference on Computer Vision (ECCV). Munich: Springer, 2018: 552-568.

[1]	卓亚琦, 魏家辉, 李志欣. 基于双注意模型的图像描述生成方法研究[J]. 电子学报, 2022, 50(5): 1123-1130.
[2]	袁冠, 邴睿, 刘肖, 代伟, 张艳梅, 蔡卓. 基于时空图神经网络的手势识别[J]. 电子学报, 2022, 50(4): 921-931.
[3]	孟琭, 徐磊, 郭嘉阳. 一种基于改进的MobileNetV2网络语义分割算法[J]. 电子学报, 2020, 48(9): 1769-1776.
[4]	张丞, 何坚, 王伟东. 空间上下文与时序特征融合的交警指挥手势识别技术[J]. 电子学报, 2020, 48(5): 966-974.
[5]	李宝奇, 贺昱曜, 强伟, 何灵蛟. 基于并行附加特征提取网络的SSD地面小目标检测模型[J]. 电子学报, 2020, 48(1): 84-91.
[6]	王勇, 王沙沙, 田增山, 周牧, 吴金君. 基于FMCW雷达的双流融合神经网络手势识别方法[J]. 电子学报, 2019, 47(7): 1408-1415.
[7]	李宝奇, 贺昱曜, 何灵蛟, 强伟. 基于全卷积神经网络的非对称并行语义分割模型[J]. 电子学报, 2019, 47(5): 1058-1064.
[8]	冯志全, 杨学文, 徐涛, 刘弘, 吕娜, 杨晓辉, 徐治鹏. 结合手势二进制编码和类-Hausdorff距离的手势识别[J]. 电子学报, 2017, 45(9): 2281-2291.
[9]	任海兵;祝远新;徐光;林学;张哓平. 基于视觉手势识别的研究—综述[J]. 电子学报, 2000, 28(2): 118-121.

基于IASPP-ResNet分割算法的手势识别

Hand Gesture Recognition Based on IASPP-ResNet Segmentation Algorithm

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 18

参考文献 40

相关文章 9

编辑推荐

Metrics