Surface Blemishes of Aluminum Material Image Recognition Based on Transfer Learning

Abstract

Image datasets in the field of industrial manufacturing usually have the problem of uneven distribution of samples. To solve this problem, this paper utilizes transfer learning to recognize surface blemishes of Aluminum material image. In order to get rid of the disadvantage of VGGNet too many parameters, this paper combines VGGNet with Network-in-Network to generate a new model. The results of experiments is shown that it has achieved significant improvement by using transfer learning. Additionally, the new model also achieves a better performance and the number of parameters of the new model is much smaller than that of the original VGGNet.