基于多尺度特征与注意力机制的宫颈病变检测

doi:10.16180/j.cnki.issn1007-7820.2024.10.005

Abstract

Abstract:

CIN(Cervical Intraepithelial Neoplasm) is a precancerous lesion of the cervix with a high correlation to invasive cervical cancer. Accurate detection and classification of CIN is helpful to reduce the rate of severe cervical cancer. YOLOv5-CBTR(You Only Look Once version 5-Convolutional Block Transformer) cervical lesion detection method is proposed to address the issues of low accuracy in detection and classification of cervical lesions by combining multi-scale features and multiple attention mechanisms. The backbone network employs the SE-CSP (SENet-BottleneckCSP) with SENet (Squeeze-and-Excitation Networks) attention mechanism for feature extraction. The Transformer encoder module is introduced to fuse and amplify multi-feature information, and multi-head attention mechanism is used to enhance the feature extraction ability of lesion regions. Convolutional attention modules are introduced into the feature fusion layer for multiscale fusion of lesion feature information. The power transformation is introduced into the calculation of the boundary regression box, which speeds up the convergence of the model's loss function and realizes the detection and classification of cervical lesions. The experimental results show that the accuracy, recall rate, mAP(mean Average Precision), and F value of YOLOv5-CBTR model for the detection and classification of RGB cervical lesion images are 93.99%, 92.91%, 92.80%, and 93.45%, respectively. The mAP and F values of the model in multispectral cervical image detection and classification are 97.68% and 95.23%, respectively.

Key words: cervical image, lesion detection, multiscale features, attention mechanism, multispectral image, transformer encoder, power transformation, deep learning

CLC Number:

TP391

FENG Ting, YING Jie, YANG Haima, LI Fang. Detection of Cervical Lesions Based on Multi-Scale Features and Attention Mechanism[J].Electronic Science and Technology, 2024, 37(10): 30-39.

Figures/Tables 21

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Table 1.

Table 2.

Table 3.

Figure 11.

Figure 12.

Figure 13.

Table 4.

Figure 14.

Table 5.

Table 6.

Figure 15.

References 18

[1]	张灿, 陈玮, 尹钟. 基于弱监督宫颈细胞图像的语义分割方法[J]. 电子科技, 2021, 34(12):68-74.
	Zhang Can, Chen Wei, Yin Zhong. Semantic segmentation of cervical cell image based on weak supervision[J]. Electronic Science and Technology, 2021, 34(12):68-74.
[2]	Freddie, Bray, Jacques, et al. Global cancer statistics:Globocan estimates of incidence and mortality woroldwide for 36 cancers in 185 countries[J]. Ca:A Cancer Journal for Clinicians, 2018, 68(6):394-424.
[3]	Chitra B, Kumar S S. Recent advancement in cervical cancer diagnosis for automated screening:A detailed review[J]. Journal of Ambient Intelligence and Humanized Computing, 2022, 13(1):251-269.
[4]	Fan J F, Liu J, Xie S, et al. Cervical lesion image enhancement based on conditional entropy generative adversarial network framework[J]. Methods, 2021, 203(5):523-532.
[5]	Rema P N, Matnew A, Thomas S. Performance of colposcopic scoring by modified international federation of cervical pathology and colposcopy terminology for diagnosing cervical intraepithelial neoplasia in a low-resource setting[J]. South Asian Journal of Cancer, 2019, 8(4):218-220. doi: 10.4103/sajc.sajc_302_18 pmid: 31807480
[6]	Chen K, Xuan Y, Lin A, et al. Esophageal cancer detection based on classification of gastrointestinal CT images using improved Faster RCNN[J]. Computer Methods and Programs in Biomedicine, 2021, 207(2):106-172.
[7]	Li Y X, Chen J W, Xue P, et al. Computer-aided cervical cancer diagnosis using time-lapsed colposcopic images[J]. IEEE Transactions on Medical Imaging, 2020, 39(11):3403-3415.
[8]	Yue Z J, Ding S, Zhao W D, et al. Automatic CIN grades prediction of sequential cervigram image using LSTM with multistate CNN features[J]. IEEE Journal of Biomedical and Health Informatics, 2020, 24(3):844-854. doi: 10.1109/JBHI.2019.2922682 pmid: 31199278
[9]	Chen T T, Liu X C, Feng R W, et al. Discriminative cervical lesion detection in colposcopic images with global class activation and local bin excitation[J]. IEEE Journal of Biomedical and Health Informatics, 2022, 26(4):1411-1421.
[10]	Bai B, Du Y Z, Liu P Z, et al. Detection of cervical lesion region from colposcopic images based on feature reselection[J]. Biomedical Signal Processing and Control, 2020, 57(2):1-8.
[11]	Chen J C, Li P, Xu T X, et al. Detection of cervical lesions in colposcopic images based on the RetinaNet method[J]. Biomedical Signal Processing and Control, 2022, 75(1):1-17.
[12]	李燕云, 王永明, 周奇, 等. 基于宫颈上皮与血管特征的阴道镜图像深度学习模型探索[J]. 复旦学报(医学版), 2021, 48(4):435-442.
	Li Yanyun, Wang Yongming, Zhou Qi, et al. Deep learning model exploration of colposcopy image based on cervical epithelial and vascular features[J]. Fudan University Journal of Medical Sciences, 2021, 48(4):435-442.
[13]	Jie H, Li S, Samuel A, et al. Squeeze-and-excitation net-works[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8):2011-2023. doi: 10.1109/TPAMI.2019.2913372 pmid: 31034408
[14]	Ashsh V, Noam S, Niki P, et al. Attention is all you need[J]. Learning, 2017, 5(1):90-105.
[15]	Woo S, Park J, Lee J Y, et al. CBAM:Convolutional block attention module[J]. Lecture Notes in Computer Science, 2018, 11(2):3-19.
[16]	Li X, Xu Z H, Shen X, et al. Detection of cervical cancer cells in whole slide images using deformable and global context aware Faster RCNN-FPN[J]. Current Oncology, 2021, 28(5):3585-3601. doi: 10.3390/curroncol28050307 pmid: 34590614
[17]	Wan J J, Chen B, Yu Y T, et al. Polyp detection from colorectum images by using attentive YOLOv5[J]. Diagnostics, 2021, 11(12):2264-2279.
[18]	何龙兵, 那彦. 多光谱可见光图像与高分辨率图像的分维融合[J]. 电子科技, 2011, 24(1):4-8.
	He Longbing, Na Yan. Fractal fusion for more spectral visible light images and high-resolution panchromatic images[J]. Electronic Science and Technology, 2011, 24(1):4-8.

模型	CIN2	CIN3	Cancer	平均值
DGCA-RCNN^[16]	73.25	78.68	81.34	77.76
Retina Net^[9]	86.25	90.53	95.02	90.60
YOLOv5^[17]	68.80	69.60	94.70	77.70
YOLOv5-CBTR	89.30	94.41	98.26	93.99

模型	CIN2	CIN3	Cancer	平均值
DGCA-RCNN^[16]	61.50	57.33	77.38	65.40
Retina Net^[9]	79.71	71.82	89.63	80.39
YOLOv5^[17]	81.00	84.10	90.90	85.30
YOLOv5-CBTR	87.60	92.61	98.52	92.91

模型	mAP	准确率	召回率	F度量
DGCA- RCNN^[16]	68.99	77.76	65.40	71.05
Retina Net^[9]	75.98	90.60	80.39	85.19
YOLOv5^[17]	86.10	77.70	85.30	81.32
YOLOv5-CBTR	92.80	93.99	92.91	93.45

数据集	mAP	准确率	召回率	F度量
RGB	92.80	93.99	92.91	93.45
Iodine stained	80.66	83.31	83.32	83.31
Multispectral	97.68	94.20	97.70	95.23

模型	CIN2	CIN3	Cancer	平均值
YOLOv5+C3TR	85.00	85.60	96.90	89.20
YOLOv5+SE-CSP	84.30	86.71	92.32	87.78
YOLOv5+CBAM	85.10	91.50	97.90	91.50
YOLOv5-CBTR	89.30	94.41	98.26	93.99

Detection of Cervical Lesions Based on Multi-Scale Features and Attention Mechanism

RichHTML

PDF (PC)

Like

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 21

References 18

Related Articles 15

Metrics

Comments

Recommended 0

模型	CIN2	CIN3	Cancer	平均值
YOLOv5+C3TR	86.70	88.50	96.60	90.60
YOLOv5+SE-CSP	86.60	87.44	90.04	88.15
YOLOv5+CBAM	87.40	89.00	96.30	90.90
YOLOv5-CBTR	87.60	92.61	98.52	92.91

[1]	KUAI Xinchen, LI Ye. Hybrid Image Super-Resolution Reconstruction with Multiple and Multi-Scale Attention [J]. Electronic Science and Technology, 2024, 37(9): 34-42.
[2]	HE Zhiqiang, SUN Zhanquan. Swin-Transformer-Based Carotid Ultrasound Image Plaque Segmentation [J]. Electronic Science and Technology, 2024, 37(9): 48-56.
[3]	HE Xing, HUANG Yongming, ZHU Yong. Pavement Pothole Detection Method Based on Improved YOLOv5 [J]. Electronic Science and Technology, 2024, 37(7): 53-59.
[4]	YE Yuxin, JU Zhiyong, LAI Ying. Traffic Sign Detection Algorithm Incorporating Receptive Field Enhancement Module and Attention Mechanism [J]. Electronic Science and Technology, 2024, 37(6): 8-16.
[5]	YUAN Ao, QI Jinpeng, JIA Can, XUE Yuxin, GUO Yangyang. A Classification Method of Time Series Pathological Data Segments Based on Deep Learning [J]. Electronic Science and Technology, 2024, 37(6): 84-91.
[6]	ZHU Zihao, SONG Yan. Lightweight Capsule Network Fusing Attention and Capsule Pooling [J]. Electronic Science and Technology, 2024, 37(5): 1-8.
[7]	XIA Rongcheng, LIU Deer. Vehicle Detection and Analysis in Urban Waterlogging Area Based on Deep Learning [J]. Electronic Science and Technology, 2024, 37(5): 18-24.
[8]	MA Wenjie, ZHANG Xuanxiong. Research on Blind Roads and Obstacle Recognition Algorithm Based on Deep Learning [J]. Electronic Science and Technology, 2024, 37(3): 75-83.
[9]	JU Ping, SONG Yan, ZHANG Yingjie, XU Yifu, SHAO Hang. Research on Agricultural Crop Diseases and Pests Classification Based on Masked Autoencoding [J]. Electronic Science and Technology, 2024, 37(10): 23-29.
[10]	ZHANG Yang, LIU Jin. Character Enhancement Based on Named Entity Recognition for Industrial Equipment Faults [J]. Electronic Science and Technology, 2024, 37(10): 48-54.
[11]	XIN Tenghao, LI Feifei. Research on Action Recognition Method Based on Deep Learning [J]. Electronic Science and Technology, 2024, 37(10): 64-70.
[12]	ZHANG Hui, LI Feifei. Emotion Recognition Algorithm Based on Multimodal Cross-Interaction [J]. Electronic Science and Technology, 2024, 37(10): 81-87.
[13]	YU Zhihong,LI Feifei. Semi-Supervised Medical Image Segmentation Method Based on Meta-Learning and Neural Architecture Search [J]. Electronic Science and Technology, 2024, 37(1): 17-23.
[14]	HUANG Zixuan,LI Qiaoxing. Research on Identification of Urban Illegal Vehicles Based on Random Forest Model [J]. Electronic Science and Technology, 2024, 37(1): 66-71.
[15]	LI Zenghui,WANG Wei. Research Progress of Medical Image Segmentation Method Based on Deep Learning [J]. Electronic Science and Technology, 2024, 37(1): 72-80.