Year: 2024
Type: conference-paper
Abstract: This research is committed to exploring how deep learning technology can be applied to the automatic classification process of financial accounting vouchers and their attachments, aiming to improve the efficiency and accuracy of classification work. The research team has designed a model based on convolutional neural network, which can automatically distinguish different types of documents, such as bookkeeping vouchers, bank statements and invoices, through in-depth learning of the scanned images of massive accounting vouchers and attachment materials. The research first combs the limitations of traditional methods and machine learning in the field of voucher classification, and then elaborates the specific steps of data collection, model construction, and training optimization. The experimental results show that the depth learning model has significant advantages in dealing with the challenges of large differences in image quality and uneven distribution of categories, and significantly improves the accuracy and stability of voucher classification. This research result not only reduces the work pressure for financial personnel, but also provides new innovative ideas for the supporting software of high-definition scanning equipment, which has far-reaching significance for promoting the digital transformation process of the financial accounting industry.
Authors: Jihong Tang, Wenjuan Xu
Institutions: Jimei University, Xiamen Institute of Technology
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4409965965