• Laser & Optoelectronics Progress
  • Vol. 60, Issue 2, 0230001 (2023)
Jinyang Liu1, Mingxin Yu1,*, Shengnan Ji2, Lianqing Zhu1..., Tao Zhang1, Jingya Ding1 and Jiabin Xia1|Show fewer author(s)
Author Affiliations
  • 1Key Laboratory of Optoelectronic Measurement Technology and Instrument, Ministry of Education, School of Instrument Science and Opto-Electronics Engineering, Beijing Information Science & Technology University, Beijing 100192, China
  • 2China North Chemical Research Academy Group Co., Ltd., Beijing 100089, China
  • show less
    DOI: 10.3788/LOP212701 Cite this Article Set citation alerts
    Jinyang Liu, Mingxin Yu, Shengnan Ji, Lianqing Zhu, Tao Zhang, Jingya Ding, Jiabin Xia. Raman Spectral Segmentation Method for Tongue Squamous Cell Carcinoma Using Deep Learning[J]. Laser & Optoelectronics Progress, 2023, 60(2): 0230001 Copy Citation Text show less
    Effect drawing of data preprocessing
    Fig. 1. Effect drawing of data preprocessing
    Average Raman characteristic peaks of tongue squamous cell carcinoma tissue and normal tissue
    Fig. 2. Average Raman characteristic peaks of tongue squamous cell carcinoma tissue and normal tissue
    Structure diagram of ISB-CNN model
    Fig. 3. Structure diagram of ISB-CNN model
    Conv Block and Identity Block
    Fig. 4. Conv Block and Identity Block
    Schematic of preseted band
    Fig. 5. Schematic of preseted band
    Functional diagram of candidate band pooling layer
    Fig. 6. Functional diagram of candidate band pooling layer
    P-R curve of tongue squamous cell carcinoma
    Fig. 7. P-R curve of tongue squamous cell carcinoma
    Labeling results of tongue squamous cell carcinoma
    Fig. 8. Labeling results of tongue squamous cell carcinoma
    Important spectral band regions predicted by the proposed model
    Fig. 9. Important spectral band regions predicted by the proposed model
    Raman shifting /cm-1Corresponding material
    798Nucleic
    993Phenylalanine
    1236Amide Ⅲ
    1376Phospholipid (CH3)
    1451Lipid
    1659Keratin
    Table 1. Characteristic peaks of Raman spectrum of tongue squamous cell carcinoma
    Hyper parameterContent
    OptimizerAdam
    Learning rate0.001
    Batch size1
    Epoch100
    Table 2. Training hyperparameters of ISB-CNN model
    Characteristic peak /cm-1Corresponding materialManual mark area /cm-1Model output area /cm-1IoU /%
    795Nucleic[765,820][767,817]90.90
    989Phenylalanine[972,1003][971,1004]93.93
    1230AmideIII[1205,1250][1207,1254]87.75
    1376Phospholipid (CH3)[1350,1402][1353,1401]92.30
    1451Lipid[1433,1470][1433,1472]94.87
    1668Keratin[1650,1683][1649,1686]91.67
    Table 3. Comparison between predicted bands of ISB-CNN model and manually labeled bands
    Jinyang Liu, Mingxin Yu, Shengnan Ji, Lianqing Zhu, Tao Zhang, Jingya Ding, Jiabin Xia. Raman Spectral Segmentation Method for Tongue Squamous Cell Carcinoma Using Deep Learning[J]. Laser & Optoelectronics Progress, 2023, 60(2): 0230001
    Download Citation