基于VGGNet-16迁移学习的男性胸肌发达程度识别模型

A Recognition Model of Male Chest Muscle Development Based on VGGNet-16 Transfer Learning

VGGNet是由牛津大学计算机视觉组和Google DeepMind公司的研究员共同研发的一例经典CNN模型,在计算机图像识别、分类工作上具有强大的优势。研究采用VGGNet-16迁移学习方式构建胸肌发达程度的识别模型,探索利用计算机深度学习技术识别肌肉发达程度的可行性。结论:基于VGGNet-16预训练模型(ImageNet)对胸肌图像进行迁移学习,模型经过30次迭代学习,最终在测试集上的准确率达到95.2%,损失达到0.0658,能够很好地提取胸肌图像特征,并能够较为准确地识别胸肌发达程度。

VGGNet is a classic CNN model developed by researchers from the computer vision group of Oxford University and Google deepmind company jointly,VGGNet has strong advantages in computer image recognition and classification.In this study,Vggnet-16 transfer learning method is used to build an evaluation model of the degree of chest muscle development and to explore the feasibility of using computer deep learning technology to evaluate the degree of muscle development.Conclusion:based on the VGGNet-16 pre training model(ImageNet),the transfer learning of the chest muscle image is carried out.After 30 times of iterative learning,the accuracy of the model in the test set is 95.2%,and the loss is 0.0658.It can extract the characteristics of the chest muscle image well and evaluate the development of the chest muscle accurately.