Negative sample labeling training method and highly automatic bill identification method
A technology of training method and identification method, applied in the field of intelligent accounting, can solve the problems of low work efficiency and high labor intensity of bill identification, and achieve the effect of improving efficiency, reducing labor participation, and preventing the difference from being too large.
Active Publication Date: 2020-09-04
西安网算数据科技有限公司
View PDF0 Cites 7 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004]
The present invention proposes a negative sample labeling training method and a highly automated bill recognition method, which solves the problems of high manual labor intensity and low work efficiency in bill recognition in the prior art
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0063] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Apparently, the described embodiments are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.
[0064] like figure 1 As shown, a negative sample labeling model training method, including
[0065] S0: Select the training sample set A of the negative sample, manually label the bills in the training sample set A of the negative sample, obtain the offset set B of the real label frame, and give a confidence level Y of the real frame according to the manual labeling result * ;
[0066] S1: Use VGG16 as the network model, and combine the pyramid feature network to extract th...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention belongs to the technical field of intelligent account making, and provides a negative sample annotation training method and a highly automatic bill identification method, which comprisesthe following steps: training a negative sample annotation model; constructing a bill warehouse D; training a bill content recognition model F through a deep learning method according to the bill pictures in the bill warehouse D; loading the bill content identification model F; performing bill identification; and inputting the bills of which the identification error times are greater than two into the negative sample labeling model, labeling again, then putting the bills into the bill warehouse D, retraining the bill content identification model F, and carrying out bill identification again.Through the method, the problems of high manual labor intensity and low working efficiency of bill identification in the prior art are solved.
Description
technical field [0001] The invention belongs to the technical field of intelligent accounting, and relates to a negative sample labeling training method and a highly automated bill recognition method. Background technique [0002] With the rapid economic development in recent years, bills are an indispensable communication "bridge". Due to the large number of bills and rich bill fields, the traditional manual entry mode cannot adapt to the pace of modern enterprise progress, so it is imminent to solve the bill entry problem. In some respects, machine vision automatic recognition entry has more advantages than manual entry. The bill recognition system can perform high-precision, high-efficiency and highly automated recognition and classification of various bill tasks based on data and image analysis. The bill recognition system can not only reduce work tasks and pressure and improve office efficiency, but also solve the contradictions caused by rising labor costs and labor s...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/20G06K9/62
CPCG06V10/22G06F18/241G06F18/214Y02T10/40
Inventor 张汉宁苏斌弋渤海杨芳
Owner 西安网算数据科技有限公司