ResNet对更年期泌尿生殖系统综合症光学相干断层成像图像的识别

Recognition of menopausal urogenital syndrome image by ResNet50

计算机的软硬件发展使得深度学习广泛地应用于生活的各方面,尤其是图像识别分类领域。在医学领域,更年期泌尿生殖系统综合症已经严重影响着妇女的健康和性功能方面的生活质量,目前通过光学相干断层成像(OCT)系统可以得到实时光学活检图像,研究运用深度学习算法可以更高效地完成患者和健康人群的图像识别,使用新型内窥镜OCT在体内对妇女阴道成像,分别以Inception-v3,VGG16,ResNet50三个网络模型结构用于健康人、患者和正在接受治疗患者的数据集,通过对比数据得到ResNet50在不影响精度下收敛迅速,消耗时间更短,因此选择ResNet50进行识别更为合适。

The development of computer software and hardware makes deep learning widely used in all aspects of life,especially in the field of image recognition and classification.In the field of medicine,menopausal urogenital syndrome has seriously affected women’s health and quality of life in terms of sexual function,At present,real-time optical biopsy images can be obtained by optical coherence tomography(OCT)system.In this paper,the deep learning algorithm can be used to more efficiently complete the image recognition of patients and healthy people.A new type of endoscopic OCT is used to image women’s vagina in vivo,and this paper proposes three network model structures of Inception-v3,VGG16 and ResNet50,which are used in the test group of healthy people,patients and patients undergoing treatment.By comparing the data,it is found that ResNet50 converges rapidly and consumes less time without affecting the accuracy,so it is more appropriate to select ResNet50 for recognition.By comparing the data,ResNet50 converges rapidly without sacrificing the accuracy and consumes less time,so it is more appropriate to select ResNet50 for recognition.