Share Email Print
cover

Proceedings Paper

An opto-electronic joint detection system based on DSP aiming at early cervical cancer screening
Author(s): Weiya Wang; Mengyu Jia; Feng Gao; Lihong Yang; Pengpeng Qu; Changping Zou; Pengxi Liu; Huijuan Zhao
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The cervical cancer screening at a pre-cancer stage is beneficial to reduce the mortality of women. An opto-electronic joint detection system based on DSP aiming at early cervical cancer screening is introduced in this paper. In this system, three electrodes alternately discharge to the cervical tissue and three light emitting diodes in different wavelengths alternately irradiate the cervical tissue. Then the relative optical reflectance and electrical voltage attenuation curve are obtained by optical and electrical detection, respectively. The system is based on DSP to attain the portable and cheap instrument. By adopting the relative reflectance and the voltage attenuation constant, the classification algorithm based on Support Vector Machine (SVM) discriminates abnormal cervical tissue from normal. We use particle swarm optimization to optimize the two key parameters of SVM, i.e. nuclear factor and cost factor. The clinical data were collected on 313 patients to build a clinical database of tissue responses under optical and electrical stimulations with the histopathologic examination as the gold standard. The classification result shows that the opto-electronic joint detection has higher total coincidence rate than separate optical detection or separate electrical detection. The sensitivity, specificity, and total coincidence rate increase with the increasing of sample numbers in the training set. The average total coincidence rate of the system can reach 85.1% compared with the histopathologic examination.

Paper Details

Date Published: 26 February 2015
PDF: 10 pages
Proc. SPIE 9303, Photonic Therapeutics and Diagnostics XI, 93032W (26 February 2015); doi: 10.1117/12.2076587
Show Author Affiliations
Weiya Wang, Tianjin Key Lab. of Biomedical Detecting Techniques and Instruments (China)
Mengyu Jia, Tianjin Key Lab. of Biomedical Detecting Techniques and Instruments (China)
Feng Gao, Tianjin Key Lab. of Biomedical Detecting Techniques and Instruments (China)
Lihong Yang, Tianjin Central Hospital of Gynecology Obstetrics (China)
Pengpeng Qu, Tianjin Central Hospital of Gynecology Obstetrics (China)
Changping Zou, Univ. of Connecticut Health Ctr. (United States)
Pengxi Liu, Tianjin Key Lab. of Biomedical Detecting Techniques and Instruments (China)
Huijuan Zhao, Tianjin Key Lab. of Biomedical Detecting Techniques and Instruments (China)


Published in SPIE Proceedings Vol. 9303:
Photonic Therapeutics and Diagnostics XI
Hyun Wook Kang; Brian J. F. Wong; Melissa C. Skala; Bernard Choi; Guillermo J. Tearney; Andreas Mandelis; Nikiforos Kollias; Kenton W. Gregory; Mark W. Dewhirst; Justus F. Ilgner; Alfred Nuttal; Haishan Zeng; Laura Marcu; Claus-Peter Richter, Editor(s)

© SPIE. Terms of Use
Back to Top