Detection of cervical lesions in colposcopic images based on the RetinaNet method. (May 2022)

• Record Type: Journal Article
• Authors: Chen, Jiancui; Li, Ping; Xu, Tianxiang; Xue, Huifeng; Wang, Xiaoxia; Li, Ye; Lin, Hao; Liu, Peizhong; Dong, Binhua; Sun, Pengming
• Is Part Of: Biomedical signal processing and control. Volume 75(2022)
• Publication Date: 2022-05
• Subjects: FPN feature pyramid network -- mAP mean average precision -- AR average recall -- AI artificial intelligence -- CAD computer-aided diagnosis -- CIN cervical intraepithelial neoplasia -- LSIL low-grade squamous intraepithelial lesions -- HSIL high-grade squamous intraepithelial lesions -- ROIs regions of interest -- PLAB pyramid histogram in LAB color space -- PHOG pyramid histogram of the oriented gradient -- PLBP pyramid histogram of the local binary pattern -- SSD single-shot multi-box detector -- FCN fully convolutional network -- RPNs region proposal networks -- CNN convolutional neural network -- GT ground truth -- ReLU rectified linear unit -- IoU intersection over union -- mAP mean average precision* -- SGD stochastic gradient descent -- LR learning rate -- YOLO you only look once -- ACC accuracy -- SEN sensitivity -- SPEC specificity -- PPV positive predictive value -- NPV negative predictive value; Cervical lesion -- Colposcopic images -- Object detection -- RetinaNet; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2022.103589
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlight: We propose a method of detecting cervical lesion area in colposcopy image based on RetinaNet method. To overcome the shortcomings of low contrast and high visual similarity with normal areas, the depth features of the lesion areas are extracted, and unimportant features are suppressed. The lesion areas in colposcopic images are further detected, and the category and risk coefficient of the lesion areas are evaluated. Abstract: There is a critical requirement for detecting cervical lesions through colposcopic images in computer-aided diagnosis. Compared to images from natural scenes, colposcopic images have some specific problems, such as low contrast, high visual similarity, and blurry lesion boundaries that make it difficult to accurately detect cervical lesion areas. To solve these problems, this paper proposes a method based on RetinaNet to detect lesion areas in colposcopic images. First, the depth features of the entire image are extracted by a fusion of ResNet50 and a feature pyramid network (FPN). In addition, the model suppresses the weight of simple and easy-to-distinguish samples through local loss, ensuring that the training can focus on the hard-to-distinguish and that important samples, while improving the utilization rate of the important features. Then, object classification and bounding box regression are performed on the feature map through two subnets. Under the same experimental conditions, the detection effects of this method are compared with those of other mainstream models through the mean average precision (mAP), average recall (AR) and other indexes. Experimental results show that the method based on RetinaNet is superior to these compared models, with a mAP [.5:.95] of 32.72%, a mAP.5 of 50.16%, and an AR of 49.70%. Compared with those of Faster R-CNN-ResNet50 + FPN, the mAP [.5:.95] is 2.76% higher and the AR is 6.42% higher.