Applications of Deep Learning in Unified Credit Management of Commercial Banks
Yang Wang, Zhendong Li, Xia-Qing Xi · 2020
With the rapid development of new technologies, such as cloud computing, mobile internet, financial technology and big data, commercial banks in China have to step up their efforts to use new technologies such as artificial intelligence to innovate. In this paper, we focus on detecting the image orientation and rotating the image to the right angle automatically based on convolutional neural network in unified credit management of banks. First, in view of the characteristics of scanned images in banking systems, we analyze the process of credit evaluation and approval of Zhengzhou bank in detail. Then, we convert the problem of skew correction into image classification in unified credit management of commercial banks, and we fine-tuning the pretrained VGG16 model with TensorFlow for image classification. Finally, we develop a deep learning platform in bank of Zhengzhou to test the performance of the proposed method. In contrast to traditional image processing methods, the test results on bank's real dataset show the superiority of our proposed method, and the proposed method has been applied in the process of credit approval to improve the automation and intelligence level of Zhengzhou Bank to a certain extent.