摘要
使用训练集的80%训练了基于ResNet-101的预测模型,剩余20%作为测试集用于评估5种出血类型的效能。实验结果表明,每一张图像的预测准确率为94.6%,每一类的平均预测准确率达98.1%。
80%training dataset is used to train the ResNet-101 based prediction model,while the other 20%is for testing the efficiency of five bleeding types.The results indicate that the prediction accuracy for each image is 94.6%,and the average prediction accuracy of each category is up to 98.1%.
作者
周长才
刘爽
王昕
ZHOU Changcai;LIU Shuang;WANG Xin(School of Computer Science and Engineering, Changchun University of Technology, Changchun 130012, China)
出处
《长春工业大学学报》
CAS
2021年第5期469-473,共5页
Journal of Changchun University of Technology
基金
吉林省教育厅科学技术研究基金资助项目(JJKH20210739KJ)。
