Biological Tissue Recognition Based on X-Ray Absorption Spectral Detection†

doi:10.7503/cjcu20170830

Abstract

Abstract:

In order to verify the ability of X-ray absorption spectroscopy to recognize biological tissues, porcine heart, liver, kidney, stomach, lean meat and fat are selected as specimens. The X-ray absorption spectra of these specimens are obtained by X-ray detector at the excitation voltage of 55 kV. The collected spectra are divided into training set and test set, and principal component analysis is used to extract spectral principal component. Using the training set as input, a radial basis function(RBF) neural network model is established to predict the samples of the test set. The recognition rate of all samples through cross-validation method reached 90.22%. The experimental results show that X-ray spectral technique coupled with statistical analysis methods can be used for the classification of pig tissue, which is of great significance to the application of X-ray spectral technique in organism identification.

Key words: X-Ray absorption spectrum, Biological tissue, Material recognition, Radial basis function(RBF) neural network

CLC Number:

O657.39

TrendMD:

WANG Qian, YANG Zheng, FANG Zheng. Biological Tissue Recognition Based on X-Ray Absorption Spectral Detection^†[J]. Chem. J. Chinese Universities, 2018, 39(7): 1434.

Figures/Tables 9

References 23

[1]	Zhang Z.X.,Acta Laser Biology Sinica, 2004, (3), 159—160
	(张镇西. 激光生物学报, 2004, (3), 159—160)
[2]	Manoharan R., Wang Y., Feld M S., Spectrochim. Acta A, 1996, 52(2), 215—249
[3]	Siqueira L. F. S., Araujo R. F., De Araujo A. A., Morais C. L. M., Lima K. M. G., Chemometr. Intell. Lab., 2017, 162, 123—129
[4]	Lambrakos S. G., Trzaskoma-Paulette P. P., Appl. Spectrosc., 2000, 54(2), 305—315
[5]	Yi W. S., Cui D. S., Li Z., Wu L. L., Shen A. G., Hu J. M., Spectrochim. Acta A, 2013, 101, 127—131
[6]	Cheng C. G., Tian Y. M., Zhang C. J., Chinese J. Anal. Chem., 2008, 36(8), 1051—1055
	(程存归,田玉梅,张长江. 分析化学,2008, 36(8), 1051—1055)
[7]	Bard M. P. L., Amelink A., Hegt V. N., Graveland W. J., Sterenborg H. J. C. M., Hoogsteden H. C., Aerts J. G. J. V., Am. J. Resp. Crit. Care, 2005, 171(10), 1178—1184
[8]	Liu G., Liu J. H., Zhang L., Yu F., Sun S. Z., Spectrosc. Spect. Anal., 2005, 25(5), 723—725
	(刘刚,刘剑虹,张林,俞帆,孙世中.光谱学与光谱分析,2008,25(5), 723—725)
[9]	Zhang H. P., Zheng C., Lu L., Zou Y. B., Wang C. H., Xu S. P., Xu W. Q., Fan Z. M., Lu L. J., Han B., Chem. J. Chinese Universities, 2014, 35(9), 1877—1882
	(张海鹏, 郑超, 路璐, 邹亚斌, 王翠花, 徐抒平, 徐蔚青, 范志民, 路来金, 韩冰.高等学校化学学报,2014, 35(9), 1877—1882)
[10]	Rehman S., Movasaghi Z., Tucker A. T., Joel S. P., Darr J. A., Ruban A. V., Rehman I. U., J. Raman Spectrosc., 2007, 38(10), 1345—1351
[11]	Cicchi R., Anand S., Crisci A., Giordano F., Rossari S., Proceeding of SPIE, 2015, 9537, 95370I
[12]	Liu H. Y., Gu W. Q., Li Q. L., Wang Y. T., Chen Z. G., Xu Q. T., Spectrosc. Spect. Anal., 2015, 35(1), 38—43
	(刘洪英, 顾文荃, 李庆利, 王依婷, 陈增淦, 徐沁同. 光谱学与光谱分析,,2015, 35(1), 38—43)
[13]	Savin S. L. P., Berko A., Blacklocks A. N., Edwards W., Chadwick A. V., C. R. Chim., 2008, 11(9), 948—963
[14]	Ma L. D., Shanghai Measurement and Testing, 2007, 34(6), 2—11
	(马礼敦. 上海计量测试,2007,34(6), 2—11)
[15]	Hu B., Zhang X. L., Ouyang Q. N., Wu X. M., Fang Z., Measurement, 2016, 93, 252—257
[16]	Yano J., Yachandra V. K., Photosynth. Res., 2009, 102, 241—254
[17]	Jolliffe I. T., Cadima J., Philos. T. R. Soc. A, 2016, 374(2065), 20150202
[18]	Schmidhuber J., Neural Networks, 2014, 61, 85—117
[19]	Schwenker F., Kestler H. A., Palm G., Neural Networks, 2001, 14(4/5), 439—458
[20]	Ramedani Z., Omid M., Keyhani A., Shamshirband S., Khoshnevisan B., Renwe. Sust. Energ. Rev., 2014, 39, 1005—1011
[21]	Fawcett T., Pattern Recogn.Lett., 2006, 27(8), 861—874
[22]	Unnikrishnan V. K., Choudhari K. S., Kulkarni S. D., Nayak R., Kartha V. B., Santhosh C., RSC Adv., 2013, 3(48), 25872—25880
[23]	Zhang Z., Han X., Pearson E., Pelizzari C., Sidky E. Y., Pan X. C., Phys. Med. Biol., 2016, 61(9), 3387—3406

Principle component	Contribution(%)	Cumulative contribution(%)	Principle component	Contribution(%)	Cumulative contribution(%)
PC1	92.388	92.388	PC6	0.251	98.892
PC2	3.737	96.126	PC7	0.186	99.078
PC3	1.570	97.696	PC8	0.160	99.238
PC4	0.590	98.286	PC9	0.124	99.362
PC5	0.355	98.641	PC10	0.105	99.467

Principle component	Contribution(%)	Cumulative contribution(%)	Principle component	Contribution(%)	Cumulative contribution(%)
PC1	92.388	92.388	PC6	0.251	98.892
PC2	3.737	96.126	PC7	0.186	99.078
PC3	1.570	97.696	PC8	0.160	99.238
PC4	0.590	98.286	PC9	0.124	99.362
PC5	0.355	98.641	PC10	0.105	99.467

No.	Sample name	TP	TN	FP	FN	Sensitivity	Specificity	PPV	NPV	Accuracy
1	Porcine fat	60	357	18	15	0.8000	0.9520	0.7692	0.9597	0.9267
2	Porcine liver	74	375	0	1	0.9867	1.0000	1.0000	0.9973	0.9978
3	Porcine kidney	56	364	11	19	0.7467	0.9707	0.8358	0.9504	0.9333
4	Porcine lean	71	370	5	4	0.9467	0.9867	0.9342	0.9893	0.9800
5	Porcine stomach	75	370	5	0	1.0000	0.9867	0.9375	1.0000	0.9889
6	Porcine heart	70	370	5	5	0.9333	0.9867	0.9333	0.9867	0.9778

No.	Sample name	TP	TN	FP	FN	Sensitivity	Specificity	PPV	NPV	Accuracy
1	Porcine fat	60	357	18	15	0.8000	0.9520	0.7692	0.9597	0.9267
2	Porcine liver	74	375	0	1	0.9867	1.0000	1.0000	0.9973	0.9978
3	Porcine kidney	56	364	11	19	0.7467	0.9707	0.8358	0.9504	0.9333
4	Porcine lean	71	370	5	4	0.9467	0.9867	0.9342	0.9893	0.9800
5	Porcine stomach	75	370	5	0	1.0000	0.9867	0.9375	1.0000	0.9889
6	Porcine heart	70	370	5	5	0.9333	0.9867	0.9333	0.9867	0.9778